diff --git a/.agents/skills/lxmf-update/SKILL.md b/.agents/skills/lxmf-update/SKILL.md
new file mode 100644
index 0000000..20396c3
--- /dev/null
+++ b/.agents/skills/lxmf-update/SKILL.md
@@ -0,0 +1,40 @@
+---
+name: lxmf-update
+description: Check upstream LXMF and RNS implementations for changes affecting the LXMF specification, run local verifier tools, and propose spec updates with citations.
+---
+
+# LXMF update skill
+
+Use this skill when asked to check whether the LXMF specification is current with upstream LXMF or RNS.
+
+## Sources of truth
+
+Prefer, in order:
+
+1. Upstream LXMF source code at a pinned release or commit.
+2. Upstream RNS source code where LXMF behavior depends on Reticulum behavior.
+3. Reproducible local traces and test vectors.
+4. Upstream prose documentation, only when consistent with source behavior.
+
+## Required workflow
+
+1. Determine the currently pinned upstream LXMF and RNS versions from `SOURCES.md`.
+2. Check whether newer upstream releases or commits exist.
+3. Identify changed files relevant to:
+   - message packing
+   - fields and flags
+   - destination handling
+   - identity/signature behavior
+   - encryption behavior
+   - propagation
+   - delivery semantics
+   - storage format if externally observable
+4. Run the verifier tools in `tools/`.
+5. Compare observed behavior against `SPEC.md`.
+6. Produce a report with:
+   - changed upstream files
+   - affected specification sections
+   - proposed edits
+   - test-vector impact
+   - unresolved ambiguities
+7. Do not modify `SPEC.md` unless explicitly asked.
diff --git a/AGENTS.md b/AGENTS.md
new file mode 100644
index 0000000..d4c9224
--- /dev/null
+++ b/AGENTS.md
@@ -0,0 +1,16 @@
+# AGENTS.md
+
+## Repository purpose
+
+This repository attempts to define an implementation-derived specification for LXMF.
+
+Do not treat upstream prose documentation as authoritative when it conflicts with observed source behavior. Prefer source code, test vectors, and reproducible traces.
+
+## Required discipline
+
+- Distinguish normative behavior from observed implementation behavior.
+- Cite exact upstream files, functions, classes, constants, and version tags.
+- Do not invent protocol rules.
+- When behavior is unclear, mark it as "undetermined" rather than guessing.
+- Prefer Perl tools for extraction, comparison, and report generation.
+- Keep generated examples reproducible from scripts under tools/.
diff --git a/ChatGPT_1.md b/ChatGPT_1.md
new file mode 100644
index 0000000..c40e197
--- /dev/null
+++ b/ChatGPT_1.md
@@ -0,0 +1,174 @@
+Yes. For Codex, the closest equivalents are:
+
+1. **`AGENTS.md`** for persistent project-wide instructions.
+2. **Codex Skills using `SKILL.md`** for reusable workflows, very similar in spirit to the Claude Code file you linked.
+3. Optionally **`PLANS.md` / ExecPlan-style documents** for long-running specification work.
+
+The important correction is that Codex does now support **Skills with `SKILL.md`**, not just `AGENTS.md`. OpenAI’s Codex documentation says a skill is a directory containing a required `SKILL.md` file, with optional `scripts/`, `references/`, `assets/`, and related support files. Codex initially sees the skill name, description, and path, then loads the full `SKILL.md` only when it decides that skill applies. ([OpenAI Developers][1])
+
+So, yes: you can do essentially the same thing that `thatSFguy` is doing, but I would structure it slightly differently for LXMF.
+
+## Recommended Codex layout
+
+For an LXMF specification repository, I would use something like:
+
+```text
+lxmf-specification/
+├── AGENTS.md
+├── SPEC.md
+├── NOTES.md
+├── SOURCES.md
+├── tools/
+│   ├── extract_lxmf_constants.pl
+│   ├── compare_upstream_lxmf.pl
+│   ├── verify_examples.pl
+│   └── make_test_vectors.pl
+├── examples/
+│   ├── lxmf_message_minimal.hex
+│   ├── lxmf_message_signed.hex
+│   └── lxmf_propagation_example.hex
+├── references/
+│   ├── lxmf_source_map.md
+│   ├── rns_dependency_map.md
+│   └── terminology.md
+└── .agents/
+    └── skills/
+        └── lxmf-update/
+            └── SKILL.md
+```
+
+## Role of each file
+
+### `AGENTS.md`
+
+Use this for always-on repository rules. Codex reads `AGENTS.md` before doing work, and the docs describe it as the mechanism for project-specific persistent guidance. ([OpenAI Developers][2])
+
+Example use:
+
+```markdown
+# AGENTS.md
+
+## Repository purpose
+
+This repository attempts to define an implementation-derived specification for LXMF.
+
+Do not treat upstream prose documentation as authoritative when it conflicts with observed source behavior. Prefer source code, test vectors, and reproducible traces.
+
+## Required discipline
+
+- Distinguish normative behavior from observed implementation behavior.
+- Cite exact upstream files, functions, classes, constants, and version tags.
+- Do not invent protocol rules.
+- When behavior is unclear, mark it as "undetermined" rather than guessing.
+- Prefer Perl tools for extraction, comparison, and report generation.
+- Keep generated examples reproducible from scripts under tools/.
+```
+
+### `.agents/skills/lxmf-update/SKILL.md`
+
+Use this for the repeatable workflow: “check upstream LXMF/RNS, compare against our spec, update citations, run verifiers, propose diffs.”
+
+That is the closest Codex analog to the Claude Code `rns-update/SKILL.md` you found.
+
+A minimal Codex skill might look like:
+
+```markdown
+---
+name: lxmf-update
+description: Check upstream LXMF and RNS implementations for changes affecting the LXMF specification, run local verifier tools, and propose spec updates with citations.
+---
+
+# LXMF update skill
+
+Use this skill when asked to check whether the LXMF specification is current with upstream LXMF or RNS.
+
+## Sources of truth
+
+Prefer, in order:
+
+1. Upstream LXMF source code at a pinned release or commit.
+2. Upstream RNS source code where LXMF behavior depends on Reticulum behavior.
+3. Reproducible local traces and test vectors.
+4. Upstream prose documentation, only when consistent with source behavior.
+
+## Required workflow
+
+1. Determine the currently pinned upstream LXMF and RNS versions from `SOURCES.md`.
+2. Check whether newer upstream releases or commits exist.
+3. Identify changed files relevant to:
+   - message packing
+   - fields and flags
+   - destination handling
+   - identity/signature behavior
+   - encryption behavior
+   - propagation
+   - delivery semantics
+   - storage format if externally observable
+4. Run the verifier tools in `tools/`.
+5. Compare observed behavior against `SPEC.md`.
+6. Produce a report with:
+   - changed upstream files
+   - affected specification sections
+   - proposed edits
+   - test-vector impact
+   - unresolved ambiguities
+7. Do not modify `SPEC.md` unless explicitly asked.
+```
+
+Codex supports skills in the CLI, IDE extension, and Codex app, and the documented repository skill location is `.agents/skills` under the current directory or repository root. ([OpenAI Developers][1])
+
+### `PLANS.md`
+
+For your use case, I would also add a planning document, because “derive a protocol specification from implementation” is not a one-shot coding task. OpenAI’s Codex cookbook describes using `PLANS.md` / ExecPlans for long-running, multi-stage work, with `AGENTS.md` telling Codex when to use it. ([OpenAI Developers][3])
+
+For LXMF, `PLANS.md` could define the required sections of each investigation:
+
+```markdown
+# LXMF Specification Investigation Plan
+
+Each investigation must include:
+
+1. Question being answered.
+2. Upstream files inspected.
+3. Exact version or commit.
+4. Relevant classes/functions/constants.
+5. Observed behavior.
+6. Whether behavior is normative, inferred, or unresolved.
+7. Test vectors needed.
+8. SPEC.md sections affected.
+9. Open questions.
+```
+
+## My recommendation
+
+For your goal, I would not start by asking Codex to “write the LXMF specification.” That is too broad and likely to produce confident prose ahead of evidence.
+
+I would instead give Codex narrow tasks such as:
+
+```text
+Using the lxmf-update skill, inspect upstream LXMF and determine the exact serialized structure of an LXMessage. Do not edit SPEC.md. Produce a report listing every field, its source-code origin, and whether its behavior is confirmed by a test vector.
+```
+
+Then:
+
+```text
+Create Perl tooling under tools/ that extracts or verifies the LXMF message packing behavior described in the report. Generate at least one reproducible test vector.
+```
+
+Then:
+
+```text
+Update SPEC.md section "LXMessage Wire Format" using only behavior confirmed by source inspection or verifier output. Mark all unverified behavior as unresolved.
+```
+
+## Bottom line
+
+Yes, you can accomplish the same kind of workflow with Codex. In fact, Codex gives you both mechanisms:
+
+`AGENTS.md` for standing repo rules, and `SKILL.md` under `.agents/skills/...` for a reusable “LXMF update / spec maintenance” workflow. OpenAI’s Codex docs explicitly describe customization as a combination of `AGENTS.md`, skills, MCP, memories, and subagents; for your use case, `AGENTS.md` plus a `SKILL.md` skill is the right starting point. ([OpenAI Developers][4])
+
+[1]: https://developers.openai.com/codex/skills "Agent Skills – Codex | OpenAI Developers"
+[2]: https://developers.openai.com/codex/guides/agents-md "Custom instructions with AGENTS.md – Codex | OpenAI Developers"
+[3]: https://developers.openai.com/cookbook/articles/codex_exec_plans "Using PLANS.md for multi-hour problem solving"
+[4]: https://developers.openai.com/codex/concepts/customization "Customization – Codex | OpenAI Developers"
+
diff --git a/Codex_1.md b/Codex_1.md
new file mode 100644
index 0000000..049676a
--- /dev/null
+++ b/Codex_1.md
@@ -0,0 +1,8 @@
+This directory: /usr/local/src/reticulum/LXMF
+contains my GitHub fork (https://github.com/jlpoolen/LXMF) of Mark Qvist's: https://github.com/markqvist/LXMF which, incidentally, is a public mirror of some git server served up through Reticulum -- the source is not provided.
+
+I created an empty project in my Forgejo server: LXMF-specification at  https://salemdata.net/repo/jlpoole/LMXF-specification
+and I added my Forgejo instance as a remote repository identified by "forgejo".
+
+I added some files based on a ChatGPT chat, see ChatGPT_1.md.
+So, for your first task:  Using the lxmf-update skill, inspect upstream LXMF and determine the exact serialized structure of an LXMessage. Do not edit SPEC.md. Produce a report listing every field, its source-code origin, and whether its behavior is confirmed by a test vector.
diff --git a/Codex_Response_1.md b/Codex_Response_1.md
new file mode 100644
index 0000000..8b45de7
--- /dev/null
+++ b/Codex_Response_1.md
@@ -0,0 +1,83 @@
+No, I don’t need `SPEC.md` or `SOURCES.md` for this first report. I used the checked-out LXMF source directly. I did not edit `SPEC.md` or any other file.
+
+**LXMessage Serialized Structure**
+
+Source inspected: LXMF commit `fab12ad9bf9f997797034950f289fe41a79dcf5a`
+
+Core `lxmf_bytes` layout:
+
+```text
+destination_hash || source_hash || signature || msgpack_payload
+```
+
+Where `msgpack_payload` is:
+
+```text
+[
+  timestamp,
+  title,
+  content,
+  fields,
+  optional_stamp
+]
+```
+
+The fifth element is only present when a normal message stamp is generated and appended.
+
+| Field | Serialized position | Type / size | Source origin | Test vector confirmed |
+|---|---:|---|---|---|
+| `destination_hash` | bytes `0..15` | 16 bytes | [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:383), length from [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:40) and RNS 128-bit truncated hashes in [Reticulum.py](/usr/local/src/reticulum/Reticulum/RNS/Reticulum.py:146) | No |
+| `source_hash` | bytes `16..31` | 16 bytes | [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:384) | No |
+| `signature` | bytes `32..95` | 64 bytes | [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:378), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:385); RNS signature length from [Identity.py](/usr/local/src/reticulum/Reticulum/RNS/Identity.py:59) and [Identity.py](/usr/local/src/reticulum/Reticulum/RNS/Identity.py:80) | No |
+| `msgpack_payload` | bytes `96..end` | msgpack list | [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:362), packed at [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:381) | No |
+| `timestamp` | payload index `0` | Python float, normally msgpack float64 on this RNS umsgpack | [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:357), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:362); float64 behavior from [umsgpack.py](/usr/local/src/reticulum/Reticulum/RNS/vendor/umsgpack.py:1177) and [umsgpack.py](/usr/local/src/reticulum/Reticulum/RNS/vendor/umsgpack.py:325) | No |
+| `title` | payload index `1` | bytes | [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:130), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:193), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:362) | No |
+| `content` | payload index `2` | bytes | [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:135), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:202), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:362) | No |
+| `fields` | payload index `3` | dict, default `{}` | [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:215), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:362) | No |
+| `stamp` | payload index `4`, optional | bytes, only if generated | [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:371), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:373); extracted on unpack at [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:754) | No |
+
+**Hash And Signature**
+
+`message_id` / `hash` is:
+
+```text
+SHA256(destination_hash || source_hash || msgpack([timestamp, title, content, fields]))
+```
+
+Source: [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:364), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:368), RNS SHA-256 at [Identity.py](/usr/local/src/reticulum/Reticulum/RNS/Identity.py:238).
+
+The signature is over:
+
+```text
+destination_hash || source_hash || msgpack([timestamp, title, content, fields]) || message_id
+```
+
+Source: [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:375), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:378), RNS signing at [Identity.py](/usr/local/src/reticulum/Reticulum/RNS/Identity.py:767).
+
+Important: the optional normal `stamp` is excluded from both `message_id` and signature. Unpack removes payload index `4` before recomputing hash/signature material: [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:754).
+
+**Outer Representations**
+
+Direct/link/resource delivery uses full `self.packed` unchanged: [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:636), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:654).
+
+Opportunistic packet delivery omits the destination hash from the packet payload because the RNS packet destination already implies it:
+
+```text
+source_hash || signature || msgpack_payload
+```
+
+Source: [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:633).
+
+Propagated and paper messages wrap/encrypt everything after the destination hash:
+
+```text
+destination_hash || encrypted(source_hash || signature || msgpack_payload)
+```
+
+Source: propagated [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:429), paper [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:449). Propagation transit additionally msgpacks `[time.time(), [lxmf_data]]`: [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:436).
+
+**Notes**
+
+The README is not authoritative here: it describes payload order as timestamp, content, title, fields, but current code serializes timestamp, title, content, fields.
+
+No test vector confirmation exists in this checkout. There is no `tools/` directory and no tracked `SPEC.md` or `SOURCES.md`, so every “Test vector confirmed” entry above is marked `No`.
diff --git a/Codex_Response_2.md b/Codex_Response_2.md
new file mode 100644
index 0000000..8b45de7
--- /dev/null
+++ b/Codex_Response_2.md
@@ -0,0 +1,83 @@
+No, I don’t need `SPEC.md` or `SOURCES.md` for this first report. I used the checked-out LXMF source directly. I did not edit `SPEC.md` or any other file.
+
+**LXMessage Serialized Structure**
+
+Source inspected: LXMF commit `fab12ad9bf9f997797034950f289fe41a79dcf5a`
+
+Core `lxmf_bytes` layout:
+
+```text
+destination_hash || source_hash || signature || msgpack_payload
+```
+
+Where `msgpack_payload` is:
+
+```text
+[
+  timestamp,
+  title,
+  content,
+  fields,
+  optional_stamp
+]
+```
+
+The fifth element is only present when a normal message stamp is generated and appended.
+
+| Field | Serialized position | Type / size | Source origin | Test vector confirmed |
+|---|---:|---|---|---|
+| `destination_hash` | bytes `0..15` | 16 bytes | [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:383), length from [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:40) and RNS 128-bit truncated hashes in [Reticulum.py](/usr/local/src/reticulum/Reticulum/RNS/Reticulum.py:146) | No |
+| `source_hash` | bytes `16..31` | 16 bytes | [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:384) | No |
+| `signature` | bytes `32..95` | 64 bytes | [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:378), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:385); RNS signature length from [Identity.py](/usr/local/src/reticulum/Reticulum/RNS/Identity.py:59) and [Identity.py](/usr/local/src/reticulum/Reticulum/RNS/Identity.py:80) | No |
+| `msgpack_payload` | bytes `96..end` | msgpack list | [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:362), packed at [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:381) | No |
+| `timestamp` | payload index `0` | Python float, normally msgpack float64 on this RNS umsgpack | [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:357), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:362); float64 behavior from [umsgpack.py](/usr/local/src/reticulum/Reticulum/RNS/vendor/umsgpack.py:1177) and [umsgpack.py](/usr/local/src/reticulum/Reticulum/RNS/vendor/umsgpack.py:325) | No |
+| `title` | payload index `1` | bytes | [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:130), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:193), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:362) | No |
+| `content` | payload index `2` | bytes | [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:135), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:202), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:362) | No |
+| `fields` | payload index `3` | dict, default `{}` | [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:215), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:362) | No |
+| `stamp` | payload index `4`, optional | bytes, only if generated | [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:371), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:373); extracted on unpack at [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:754) | No |
+
+**Hash And Signature**
+
+`message_id` / `hash` is:
+
+```text
+SHA256(destination_hash || source_hash || msgpack([timestamp, title, content, fields]))
+```
+
+Source: [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:364), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:368), RNS SHA-256 at [Identity.py](/usr/local/src/reticulum/Reticulum/RNS/Identity.py:238).
+
+The signature is over:
+
+```text
+destination_hash || source_hash || msgpack([timestamp, title, content, fields]) || message_id
+```
+
+Source: [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:375), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:378), RNS signing at [Identity.py](/usr/local/src/reticulum/Reticulum/RNS/Identity.py:767).
+
+Important: the optional normal `stamp` is excluded from both `message_id` and signature. Unpack removes payload index `4` before recomputing hash/signature material: [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:754).
+
+**Outer Representations**
+
+Direct/link/resource delivery uses full `self.packed` unchanged: [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:636), [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:654).
+
+Opportunistic packet delivery omits the destination hash from the packet payload because the RNS packet destination already implies it:
+
+```text
+source_hash || signature || msgpack_payload
+```
+
+Source: [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:633).
+
+Propagated and paper messages wrap/encrypt everything after the destination hash:
+
+```text
+destination_hash || encrypted(source_hash || signature || msgpack_payload)
+```
+
+Source: propagated [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:429), paper [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:449). Propagation transit additionally msgpacks `[time.time(), [lxmf_data]]`: [LXMessage.py](/usr/local/src/reticulum/LXMF/LXMF/LXMessage.py:436).
+
+**Notes**
+
+The README is not authoritative here: it describes payload order as timestamp, content, title, fields, but current code serializes timestamp, title, content, fields.
+
+No test vector confirmation exists in this checkout. There is no `tools/` directory and no tracked `SPEC.md` or `SOURCES.md`, so every “Test vector confirmed” entry above is marked `No`.
diff --git a/Codex_Response_3.md b/Codex_Response_3.md
new file mode 100644
index 0000000..6362cd8
--- /dev/null
+++ b/Codex_Response_3.md
@@ -0,0 +1,27 @@
+Task #2 completed in `/usr/local/src/reticulum/LXMF`.
+
+Created:
+
+- `tools/make_test_vectors.pl`
+- `tools/verify_examples.pl`
+- `tools/LXMF/MessagePack.pm`
+- `tools/README.md`
+- `examples/lxmf_message_minimal.hex`
+- `examples/lxmf_message_stamped.hex`
+- `examples/lxmf_message_vectors.json`
+
+The dependency-free Perl tooling verifies:
+
+- Fixed hash/signature positions
+- Payload order and MessagePack types
+- Message ID calculation
+- Optional stamp exclusion from message ID and signature input
+
+Validation passed:
+
+```text
+Reticulum umsgpack byte-for-byte cross-check passed
+All LXMessage test-vector checks passed.
+```
+
+The vectors regenerate deterministically. Signature bytes are documented placeholders; Ed25519 validation is not claimed. `SPEC.md` was not created or modified.
diff --git a/PLANS.md b/PLANS.md
new file mode 100644
index 0000000..cbaedc0
--- /dev/null
+++ b/PLANS.md
@@ -0,0 +1,13 @@
+# LXMF Specification Investigation Plan
+
+Each investigation must include:
+
+1. Question being answered.
+2. Upstream files inspected.
+3. Exact version or commit.
+4. Relevant classes/functions/constants.
+5. Observed behavior.
+6. Whether behavior is normative, inferred, or unresolved.
+7. Test vectors needed.
+8. SPEC.md sections affected.
+9. Open questions.
diff --git a/examples/lxmf_message_minimal.hex b/examples/lxmf_message_minimal.hex
new file mode 100644
index 0000000..85f8d89
--- /dev/null
+++ b/examples/lxmf_message_minimal.hex
@@ -0,0 +1 @@
+000102030405060708090a0b0c0d0e0f101112131415161718191a1b1c1d1e1f202122232425262728292a2b2c2d2e2f303132333435363738393a3b3c3d3e3f404142434445464748494a4b4c4d4e4f505152535455565758595a5b5c5d5e5f94cb41d954fc40100000c40a54657374207469746c65c41744657465726d696e6973746963204c584d4620626f647980
diff --git a/examples/lxmf_message_stamped.hex b/examples/lxmf_message_stamped.hex
new file mode 100644
index 0000000..be6fc07
--- /dev/null
+++ b/examples/lxmf_message_stamped.hex
@@ -0,0 +1 @@
+000102030405060708090a0b0c0d0e0f101112131415161718191a1b1c1d1e1f202122232425262728292a2b2c2d2e2f303132333435363738393a3b3c3d3e3f404142434445464748494a4b4c4d4e4f505152535455565758595a5b5c5d5e5f95cb41d954fc40100000c40a54657374207469746c65c41744657465726d696e6973746963204c584d4620626f647980c420a0a1a2a3a4a5a6a7a8a9aaabacadaeafb0b1b2b3b4b5b6b7b8b9babbbcbdbebf
diff --git a/examples/lxmf_message_vectors.json b/examples/lxmf_message_vectors.json
new file mode 100644
index 0000000..b60d155
--- /dev/null
+++ b/examples/lxmf_message_vectors.json
@@ -0,0 +1,26 @@
+{
+   "content_hex" : "44657465726d696e6973746963204c584d4620626f6479",
+   "description" : "Deterministic LXMessage packing vectors; signature bytes are placeholders",
+   "destination_hash_hex" : "000102030405060708090a0b0c0d0e0f",
+   "fields" : {},
+   "format" : "destination_hash || source_hash || signature || msgpack_payload",
+   "message_id_hex" : "55f497f6f2ea5d858eaff541a9fa77f9d5bdf57be75be4a94df65d1f9a7ffea3",
+   "minimal_lxmf_hex" : "000102030405060708090a0b0c0d0e0f101112131415161718191a1b1c1d1e1f202122232425262728292a2b2c2d2e2f303132333435363738393a3b3c3d3e3f404142434445464748494a4b4c4d4e4f505152535455565758595a5b5c5d5e5f94cb41d954fc40100000c40a54657374207469746c65c41744657465726d696e6973746963204c584d4620626f647980",
+   "payload_order" : [
+      "timestamp",
+      "title",
+      "content",
+      "fields",
+      "optional_stamp"
+   ],
+   "payload_with_stamp_hex" : "95cb41d954fc40100000c40a54657374207469746c65c41744657465726d696e6973746963204c584d4620626f647980c420a0a1a2a3a4a5a6a7a8a9aaabacadaeafb0b1b2b3b4b5b6b7b8b9babbbcbdbebf",
+   "payload_without_stamp_hex" : "94cb41d954fc40100000c40a54657374207469746c65c41744657465726d696e6973746963204c584d4620626f647980",
+   "signature_hex" : "202122232425262728292a2b2c2d2e2f303132333435363738393a3b3c3d3e3f404142434445464748494a4b4c4d4e4f505152535455565758595a5b5c5d5e5f",
+   "signature_valid" : false,
+   "signed_part_hex" : "000102030405060708090a0b0c0d0e0f101112131415161718191a1b1c1d1e1f94cb41d954fc40100000c40a54657374207469746c65c41744657465726d696e6973746963204c584d4620626f64798055f497f6f2ea5d858eaff541a9fa77f9d5bdf57be75be4a94df65d1f9a7ffea3",
+   "source_hash_hex" : "101112131415161718191a1b1c1d1e1f",
+   "stamp_hex" : "a0a1a2a3a4a5a6a7a8a9aaabacadaeafb0b1b2b3b4b5b6b7b8b9babbbcbdbebf",
+   "stamped_lxmf_hex" : "000102030405060708090a0b0c0d0e0f101112131415161718191a1b1c1d1e1f202122232425262728292a2b2c2d2e2f303132333435363738393a3b3c3d3e3f404142434445464748494a4b4c4d4e4f505152535455565758595a5b5c5d5e5f95cb41d954fc40100000c40a54657374207469746c65c41744657465726d696e6973746963204c584d4620626f647980c420a0a1a2a3a4a5a6a7a8a9aaabacadaeafb0b1b2b3b4b5b6b7b8b9babbbcbdbebf",
+   "timestamp" : 1700000000.25,
+   "title_hex" : "54657374207469746c65"
+}
diff --git a/grep_results.txt b/grep_results.txt
new file mode 100644
index 0000000..52ae092
--- /dev/null
+++ b/grep_results.txt
@@ -0,0 +1,39 @@
+./LXMF/Utilities/lxmd.py:DEFFERED_JOBS_DELAY = 10
+./LXMF/Utilities/lxmd.py:def create_default_config(configpath):
+./LXMF/Utilities/lxmd.py:def apply_config():
+./LXMF/Utilities/lxmd.py:def lxmf_delivery(lxm):
+./LXMF/Utilities/lxmd.py:def program_setup(configdir = None, rnsconfigdir = None, run_pn = False, on_inbound = None, verbosity = 0, quietness = 0, service = False):
+./LXMF/Utilities/lxmd.py:def jobs():
+./LXMF/Utilities/lxmd.py:def deferred_start_jobs():
+./LXMF/Utilities/lxmd.py:def _request_sync(identity, destination_hash, remote_identity, timeout=15, exit_on_fail=False):
+./LXMF/Utilities/lxmd.py:def request_sync(target, remote=None, configdir=None, rnsconfigdir=None, verbosity=0, quietness=0, timeout=15, identity_path=None):
+./LXMF/Utilities/lxmd.py:def _request_unpeer(identity, destination_hash, remote_identity, timeout=15, exit_on_fail=False):
+./LXMF/Utilities/lxmd.py:def request_unpeer(target, remote=None, configdir=None, rnsconfigdir=None, verbosity=0, quietness=0, timeout=15, identity_path=None):
+./LXMF/Utilities/lxmd.py:def query_status(identity, remote_identity=None, timeout=5, exit_on_fail=False):
+./LXMF/Utilities/lxmd.py:def get_status(remote=None, configdir=None, rnsconfigdir=None, verbosity=0, quietness=0, timeout=5,
+./LXMF/Utilities/lxmd.py:def _get_target_identity(remote=None, timeout=5):
+./LXMF/Utilities/lxmd.py:def _remote_init(configdir=None, rnsconfigdir=None, verbosity=0, quietness=0, identity_path=None):
+./LXMF/Utilities/lxmd.py:def main():
+./LXMF/LXMF.py:def display_name_from_app_data(app_data=None):
+./LXMF/LXMF.py:def stamp_cost_from_app_data(app_data=None):
+./LXMF/LXMF.py:def compression_support_from_app_data(app_data=None):
+./LXMF/LXMF.py:def pn_name_from_app_data(app_data=None):
+./LXMF/LXMF.py:def pn_stamp_cost_from_app_data(app_data=None):
+./LXMF/LXMF.py:def pn_announce_data_is_valid(data):
+./LXMF/LXStamper.py:def worker_context(ctx, stamp_cost, workblock, message_id):
+./LXMF/LXStamper.py:def stamp_workblock(material, expand_rounds=WORKBLOCK_EXPAND_ROUNDS):
+./LXMF/LXStamper.py:def stamp_value(workblock, stamp):
+./LXMF/LXStamper.py:def stamp_valid(stamp, target_cost, workblock):
+./LXMF/LXStamper.py:def validate_peering_key(peering_id, peering_key, target_cost):
+./LXMF/LXStamper.py:def validate_pn_stamp(transient_data, target_cost):
+./LXMF/LXStamper.py:def validate_pn_stamps_job_simple(transient_list, target_cost):
+./LXMF/LXStamper.py:def validate_pn_stamps_job_multip(transient_list, target_cost):
+./LXMF/LXStamper.py:def validate_pn_stamps(transient_list, target_cost):
+./LXMF/LXStamper.py:def generate_stamp(message_id, stamp_cost, expand_rounds=WORKBLOCK_EXPAND_ROUNDS):
+./LXMF/LXStamper.py:def cancel_work(message_id):
+./LXMF/LXStamper.py:def job_simple(stamp_cost, workblock, message_id):
+./LXMF/LXStamper.py:def job_linux_managed(stamp_cost, workblock, message_id):
+./LXMF/LXStamper.py:def job_linux(stamp_cost, workblock, message_id):
+./LXMF/LXStamper.py:def job_android(stamp_cost, workblock, message_id):
+./docs/example_receiver.py:def delivery_callback(message):
+./docs/example_sender.py:def delivery_callback(message):
diff --git a/tools/LXMF/MessagePack.pm b/tools/LXMF/MessagePack.pm
new file mode 100644
index 0000000..5ce8c39
--- /dev/null
+++ b/tools/LXMF/MessagePack.pm
@@ -0,0 +1,181 @@
+package LXMF::MessagePack;
+
+use strict;
+use warnings;
+use Exporter 'import';
+
+our @EXPORT_OK = qw(
+    pack_array
+    pack_bin
+    pack_float64
+    pack_map
+    parse_item
+);
+
+sub pack_float64 {
+    my ($value) = @_;
+    return "\xcb" . pack("d>", $value);
+}
+
+sub pack_bin {
+    my ($value) = @_;
+    my $length = length($value);
+
+    return "\xc4" . pack("C", $length) . $value if $length <= 0xff;
+    return "\xc5" . pack("n", $length) . $value if $length <= 0xffff;
+    return "\xc6" . pack("N", $length) . $value;
+}
+
+sub pack_array {
+    my (@items) = @_;
+    my $length = scalar @items;
+    my $header;
+
+    if ($length <= 15) {
+        $header = pack("C", 0x90 | $length);
+    }
+    elsif ($length <= 0xffff) {
+        $header = "\xdc" . pack("n", $length);
+    }
+    else {
+        $header = "\xdd" . pack("N", $length);
+    }
+
+    return $header . join("", @items);
+}
+
+sub pack_map {
+    my (@entries) = @_;
+    die "pack_map requires key/value byte-string pairs\n" if @entries % 2;
+
+    my $length = @entries / 2;
+    my $header;
+
+    if ($length <= 15) {
+        $header = pack("C", 0x80 | $length);
+    }
+    elsif ($length <= 0xffff) {
+        $header = "\xde" . pack("n", $length);
+    }
+    else {
+        $header = "\xdf" . pack("N", $length);
+    }
+
+    return $header . join("", @entries);
+}
+
+sub _take {
+    my ($bytes, $offset_ref, $length) = @_;
+    die "truncated MessagePack item\n" if $$offset_ref + $length > length($bytes);
+    my $value = substr($bytes, $$offset_ref, $length);
+    $$offset_ref += $length;
+    return $value;
+}
+
+sub _parse_length {
+    my ($bytes, $offset_ref, $width) = @_;
+    my $raw = _take($bytes, $offset_ref, $width);
+    return unpack("C", $raw) if $width == 1;
+    return unpack("n", $raw) if $width == 2;
+    return unpack("N", $raw);
+}
+
+sub parse_item {
+    my ($bytes, $offset_ref) = @_;
+    $offset_ref //= \my $offset;
+    $$offset_ref //= 0;
+
+    my $start = $$offset_ref;
+    my $code = unpack("C", _take($bytes, $offset_ref, 1));
+    my $item = { code => $code };
+
+    if ($code <= 0x7f) {
+        @$item{qw(type value)} = ("integer", $code);
+    }
+    elsif ($code >= 0xe0) {
+        @$item{qw(type value)} = ("integer", $code - 256);
+    }
+    elsif (($code & 0xf0) == 0x90) {
+        _parse_array($bytes, $offset_ref, $item, $code & 0x0f);
+    }
+    elsif (($code & 0xf0) == 0x80) {
+        _parse_map($bytes, $offset_ref, $item, $code & 0x0f);
+    }
+    elsif (($code & 0xe0) == 0xa0) {
+        my $length = $code & 0x1f;
+        @$item{qw(type value)} = ("string", _take($bytes, $offset_ref, $length));
+    }
+    elsif ($code == 0xc0) {
+        $item->{type} = "nil";
+        $item->{value} = undef;
+    }
+    elsif ($code == 0xc2 || $code == 0xc3) {
+        @$item{qw(type value)} = ("boolean", $code == 0xc3 ? 1 : 0);
+    }
+    elsif ($code == 0xca) {
+        @$item{qw(type value)} = ("float32", unpack("f>", _take($bytes, $offset_ref, 4)));
+    }
+    elsif ($code == 0xcb) {
+        @$item{qw(type value)} = ("float64", unpack("d>", _take($bytes, $offset_ref, 8)));
+    }
+    elsif ($code >= 0xc4 && $code <= 0xc6) {
+        my $width = 1 << ($code - 0xc4);
+        my $length = _parse_length($bytes, $offset_ref, $width);
+        @$item{qw(type value)} = ("binary", _take($bytes, $offset_ref, $length));
+    }
+    elsif ($code >= 0xcc && $code <= 0xcf) {
+        my @formats = ("C", "n", "N", "Q>");
+        my @widths = (1, 2, 4, 8);
+        my $index = $code - 0xcc;
+        @$item{qw(type value)} = (
+            "integer",
+            unpack($formats[$index], _take($bytes, $offset_ref, $widths[$index])),
+        );
+    }
+    elsif ($code >= 0xd0 && $code <= 0xd3) {
+        my @formats = ("c", "s>", "l>", "q>");
+        my @widths = (1, 2, 4, 8);
+        my $index = $code - 0xd0;
+        @$item{qw(type value)} = (
+            "integer",
+            unpack($formats[$index], _take($bytes, $offset_ref, $widths[$index])),
+        );
+    }
+    elsif ($code == 0xd9 || $code == 0xda || $code == 0xdb) {
+        my $width = 1 << ($code - 0xd9);
+        my $length = _parse_length($bytes, $offset_ref, $width);
+        @$item{qw(type value)} = ("string", _take($bytes, $offset_ref, $length));
+    }
+    elsif ($code == 0xdc || $code == 0xdd) {
+        my $width = $code == 0xdc ? 2 : 4;
+        _parse_array($bytes, $offset_ref, $item, _parse_length($bytes, $offset_ref, $width));
+    }
+    elsif ($code == 0xde || $code == 0xdf) {
+        my $width = $code == 0xde ? 2 : 4;
+        _parse_map($bytes, $offset_ref, $item, _parse_length($bytes, $offset_ref, $width));
+    }
+    else {
+        die sprintf("unsupported MessagePack code 0x%02x at offset %d\n", $code, $start);
+    }
+
+    $item->{start} = $start;
+    $item->{end} = $$offset_ref;
+    $item->{raw} = substr($bytes, $start, $$offset_ref - $start);
+    return $item;
+}
+
+sub _parse_array {
+    my ($bytes, $offset_ref, $item, $length) = @_;
+    $item->{type} = "array";
+    $item->{items} = [map { parse_item($bytes, $offset_ref) } 1 .. $length];
+}
+
+sub _parse_map {
+    my ($bytes, $offset_ref, $item, $length) = @_;
+    $item->{type} = "map";
+    $item->{entries} = [
+        map { [parse_item($bytes, $offset_ref), parse_item($bytes, $offset_ref)] } 1 .. $length
+    ];
+}
+
+1;
diff --git a/tools/README.md b/tools/README.md
new file mode 100644
index 0000000..75538fa
--- /dev/null
+++ b/tools/README.md
@@ -0,0 +1,32 @@
+# LXMessage Test-Vector Tooling
+
+The scripts in this directory verify the implementation-derived LXMessage
+packing behavior described in `Code_Response_1.md`.
+
+They use only Perl core modules. The local `LXMF::MessagePack` module implements
+the MessagePack subset needed to generate the vectors and parse their complete
+payloads.
+
+Generate deterministic vectors:
+
+```sh
+perl tools/make_test_vectors.pl
+```
+
+Verify the vectors:
+
+```sh
+perl tools/verify_examples.pl
+```
+
+The vectors confirm:
+
+- the 16-byte destination hash, 16-byte source hash, and 64-byte signature
+  positions;
+- payload order `[timestamp, title, content, fields, optional_stamp]`;
+- float64 timestamp and binary title/content/stamp encodings;
+- the message ID calculation;
+- exclusion of the optional stamp from the message ID and signature input.
+
+The signature bytes are deterministic placeholders. These vectors do not claim
+to verify Ed25519 signing or signature validation.
diff --git a/tools/make_test_vectors.pl b/tools/make_test_vectors.pl
new file mode 100755
index 0000000..8d4385b
--- /dev/null
+++ b/tools/make_test_vectors.pl
@@ -0,0 +1,76 @@
+#!/usr/bin/env perl
+
+use strict;
+use warnings;
+use Digest::SHA qw(sha256);
+use FindBin qw($Bin);
+use File::Path qw(make_path);
+use JSON::PP;
+use lib "$Bin";
+use LXMF::MessagePack qw(pack_array pack_bin pack_float64 pack_map);
+
+my $output_dir = "$Bin/../examples";
+make_path($output_dir);
+
+my $destination_hash = pack("C*", 0x00 .. 0x0f);
+my $source_hash = pack("C*", 0x10 .. 0x1f);
+my $signature = pack("C*", 0x20 .. 0x5f);
+my $stamp = pack("C*", 0xa0 .. 0xbf);
+
+my $timestamp = 1_700_000_000.25;
+my $title = "Test title";
+my $content = "Deterministic LXMF body";
+my $fields = pack_map();
+
+my @required_items = (
+    pack_float64($timestamp),
+    pack_bin($title),
+    pack_bin($content),
+    $fields,
+);
+
+my $payload = pack_array(@required_items);
+my $stamped_payload = pack_array(@required_items, pack_bin($stamp));
+my $message_id = sha256($destination_hash . $source_hash . $payload);
+my $signed_part = $destination_hash . $source_hash . $payload . $message_id;
+
+my $minimal = $destination_hash . $source_hash . $signature . $payload;
+my $stamped = $destination_hash . $source_hash . $signature . $stamped_payload;
+
+write_hex("$output_dir/lxmf_message_minimal.hex", $minimal);
+write_hex("$output_dir/lxmf_message_stamped.hex", $stamped);
+
+my $manifest = {
+    description => "Deterministic LXMessage packing vectors; signature bytes are placeholders",
+    format => "destination_hash || source_hash || signature || msgpack_payload",
+    payload_order => [qw(timestamp title content fields optional_stamp)],
+    destination_hash_hex => unpack("H*", $destination_hash),
+    source_hash_hex => unpack("H*", $source_hash),
+    signature_hex => unpack("H*", $signature),
+    signature_valid => JSON::PP::false,
+    timestamp => $timestamp,
+    title_hex => unpack("H*", $title),
+    content_hex => unpack("H*", $content),
+    fields => {},
+    stamp_hex => unpack("H*", $stamp),
+    payload_without_stamp_hex => unpack("H*", $payload),
+    payload_with_stamp_hex => unpack("H*", $stamped_payload),
+    message_id_hex => unpack("H*", $message_id),
+    signed_part_hex => unpack("H*", $signed_part),
+    minimal_lxmf_hex => unpack("H*", $minimal),
+    stamped_lxmf_hex => unpack("H*", $stamped),
+};
+
+open my $json_fh, ">:raw", "$output_dir/lxmf_message_vectors.json"
+    or die "cannot write vector manifest: $!\n";
+print {$json_fh} JSON::PP->new->canonical->pretty->encode($manifest);
+close $json_fh;
+
+print "Generated deterministic LXMessage vectors in $output_dir\n";
+
+sub write_hex {
+    my ($path, $bytes) = @_;
+    open my $fh, ">:raw", $path or die "cannot write $path: $!\n";
+    print {$fh} unpack("H*", $bytes), "\n";
+    close $fh;
+}
diff --git a/tools/verify_examples.pl b/tools/verify_examples.pl
new file mode 100755
index 0000000..0f86e67
--- /dev/null
+++ b/tools/verify_examples.pl
@@ -0,0 +1,88 @@
+#!/usr/bin/env perl
+
+use strict;
+use warnings;
+use Digest::SHA qw(sha256);
+use FindBin qw($Bin);
+use JSON::PP;
+use lib "$Bin";
+use LXMF::MessagePack qw(pack_array parse_item);
+
+my $examples_dir = "$Bin/../examples";
+my $manifest = read_json("$examples_dir/lxmf_message_vectors.json");
+my $minimal = read_hex("$examples_dir/lxmf_message_minimal.hex");
+my $stamped = read_hex("$examples_dir/lxmf_message_stamped.hex");
+
+verify_lxmessage("minimal", $minimal, $manifest, 4);
+verify_lxmessage("stamped", $stamped, $manifest, 5);
+
+check(unpack("H*", $minimal) eq $manifest->{minimal_lxmf_hex}, "minimal full bytes match manifest");
+check(unpack("H*", $stamped) eq $manifest->{stamped_lxmf_hex}, "stamped full bytes match manifest");
+
+print "All LXMessage test-vector checks passed.\n";
+
+sub verify_lxmessage {
+    my ($name, $bytes, $expected, $expected_count) = @_;
+
+    check(length($bytes) > 96, "$name vector has fixed prefix and payload");
+    my $destination_hash = substr($bytes, 0, 16);
+    my $source_hash = substr($bytes, 16, 16);
+    my $signature = substr($bytes, 32, 64);
+    my $packed_payload = substr($bytes, 96);
+
+    check(unpack("H*", $destination_hash) eq $expected->{destination_hash_hex}, "$name destination hash");
+    check(unpack("H*", $source_hash) eq $expected->{source_hash_hex}, "$name source hash");
+    check(unpack("H*", $signature) eq $expected->{signature_hex}, "$name signature position and length");
+
+    my $offset = 0;
+    my $payload = parse_item($packed_payload, \$offset);
+    check($offset == length($packed_payload), "$name payload consumes all remaining bytes");
+    check($payload->{type} eq "array", "$name payload is a MessagePack array");
+    check(@{$payload->{items}} == $expected_count, "$name payload has $expected_count entries");
+
+    my @items = @{$payload->{items}};
+    check($items[0]{type} eq "float64", "$name payload[0] timestamp is float64");
+    check($items[0]{value} == $expected->{timestamp}, "$name payload[0] timestamp value");
+    check($items[1]{type} eq "binary", "$name payload[1] title is binary");
+    check(unpack("H*", $items[1]{value}) eq $expected->{title_hex}, "$name payload[1] title value");
+    check($items[2]{type} eq "binary", "$name payload[2] content is binary");
+    check(unpack("H*", $items[2]{value}) eq $expected->{content_hex}, "$name payload[2] content value");
+    check($items[3]{type} eq "map", "$name payload[3] fields is a map");
+    check(@{$items[3]{entries}} == 0, "$name payload[3] fields map is empty");
+
+    my $unstamped_payload = pack_array(map { $_->{raw} } @items[0 .. 3]);
+    my $message_id = sha256($destination_hash . $source_hash . $unstamped_payload);
+    my $signed_part = $destination_hash . $source_hash . $unstamped_payload . $message_id;
+
+    check(unpack("H*", $unstamped_payload) eq $expected->{payload_without_stamp_hex}, "$name canonical unstamped payload");
+    check(unpack("H*", $message_id) eq $expected->{message_id_hex}, "$name message ID excludes optional stamp");
+    check(unpack("H*", $signed_part) eq $expected->{signed_part_hex}, "$name signature input excludes optional stamp");
+
+    if ($expected_count == 5) {
+        check($items[4]{type} eq "binary", "$name payload[4] stamp is binary");
+        check(unpack("H*", $items[4]{value}) eq $expected->{stamp_hex}, "$name payload[4] stamp value");
+    }
+}
+
+sub read_json {
+    my ($path) = @_;
+    open my $fh, "<:raw", $path or die "cannot read $path: $!\n";
+    local $/;
+    return JSON::PP->new->decode(<$fh>);
+}
+
+sub read_hex {
+    my ($path) = @_;
+    open my $fh, "<:raw", $path or die "cannot read $path: $!\n";
+    local $/;
+    my $hex = <$fh>;
+    $hex =~ s/\s+//g;
+    die "$path contains non-hexadecimal data\n" if $hex =~ /[^0-9a-fA-F]/;
+    return pack("H*", $hex);
+}
+
+sub check {
+    my ($condition, $description) = @_;
+    die "not ok - $description\n" unless $condition;
+    print "ok - $description\n";
+}