feat: add reusable inbox skill bundle

AGENTS.md

@@ -32,3 +32,14 @@ For `docs/tests/inbox/`:
 - organize by folder plus `README.md`
 - do not use numeric test IDs
 - use stable case slugs and keep the roadmap synchronized with the actual files on disk
+
+## Project Skills
+
+For `skills/`:
+
+- treat this directory as project-local Codex skill packages, not general documentation
+- use skills to package reusable agent workflows, constraints, and bundled assets so other agents can use them without rediscovering repository context
+- keep each skill self-contained around `SKILL.md`; add `agents/openai.yaml` and bundled assets only when they directly support real invocation
+- if a skill depends on an executable or other runtime asset, prefer bundling the required artifact with the skill instead of relying on ad hoc rebuild instructions as the primary workflow
+- when updating an existing skill, keep `SKILL.md`, `agents/openai.yaml`, and bundled assets consistent with each other
+- if you add or materially change a project skill, update [docs/implementation-roadmap.md](/home/kurihada/project/ai-workflow-skill/docs/implementation-roadmap.md) in the same change

docs/implementation-roadmap.md

@@ -21,6 +21,7 @@ As of now:
 - `inbox` supports blocking waits, lease renewal, unread fetches backed by per-agent read cursors, `--body-file`, artifact attachments, and structured JSON errors with stable exit codes
 - integration tests now cover each implemented inbox command, plus the main inbox workflows, wait/watch flows, artifact persistence, unread behavior, and JSON error contracts
 - a human-readable inbox command test-plan set has been authored under `docs/tests/inbox/`
+- a reusable Codex skill package for `inbox` now exists under `skills/inbox/`, with a formal `SKILL.md`, `agents/openai.yaml`, and a bundled CLI binary asset
 - `orch` currently exists as a command skeleton only
 - no scheduler workflows have been implemented yet
 

skills/inbox/SKILL.md

@@ -0,0 +1,60 @@
+---
+name: inbox
+description: Durable agent-to-agent coordination through a bundled inbox CLI. Use when an agent needs to fetch work, claim a thread, report progress, ask blocked questions, wait for replies, inspect thread history, or return done/fail results through a SQLite-backed inbox instead of ad hoc chat.
+---
+
+# Inbox
+
+Use the bundled `./assets/inbox` CLI to communicate through a durable SQLite-backed inbox.
+
+## Quick Start
+
+- Invoke `./assets/inbox` relative to this skill directory.
+- Pass `--db` explicitly for every command.
+- Prefer `--json` whenever another agent or script will read the output.
+- Run `init` before first use on a new database path.
+
+## Rules
+
+- Treat `fetch` as discovery only; only `claim` grants ownership.
+- Do not start work without a valid lease.
+- Use `update --status in_progress` for progress and `update --status blocked` for precise questions.
+- Prefer `wait-reply` over manual sleep loops when blocked.
+- Use `done` for terminal success and `fail` for terminal failure.
+- Keep scheduling and dependency decisions out of this layer; those belong to `orch`.
+
+## Typical Commands
+
+```bash
+./assets/inbox --db ./coord.db --json init
+./assets/inbox --db ./coord.db --json fetch --agent backend-worker --status pending
+./assets/inbox --db ./coord.db --json claim --agent backend-worker --thread thr_123 --lease-seconds 900
+./assets/inbox --db ./coord.db --json update --agent backend-worker --thread thr_123 --status in_progress --summary "Implementing post CRUD routes"
+./assets/inbox --db ./coord.db --json update --agent backend-worker --thread thr_123 --status blocked --summary "Need auth decision" --payload-json '{"question":"Should admin auth use email/password in MVP?"}'
+./assets/inbox --db ./coord.db --json wait-reply --agent backend-worker --thread thr_123 --after-event 51 --timeout-seconds 1800
+./assets/inbox --db ./coord.db --json reply --from leader --to backend-worker --thread thr_123 --summary "Use email/password for MVP" --body "Use a simple credential flow for the first iteration."
+./assets/inbox --db ./coord.db --json done --agent backend-worker --thread thr_123 --summary "Post CRUD implemented" --body-file result.md
+```
+
+## Command Map
+
+- `init`: initialize schema and pragmas on a database path
+- `send`: create a new thread or append a message
+- `fetch`: list candidate threads for an agent without claiming them
+- `claim`: acquire a lease on a thread
+- `renew`: extend an active lease
+- `update`: send progress or blocked updates
+- `reply`: send an answer, control message, or follow-up inside a thread
+- `done`: mark a thread complete with final result data
+- `fail`: mark a thread failed with failure details
+- `cancel`: cancel a thread
+- `list`: list threads by filters
+- `show`: inspect thread state and message history
+- `watch`: block until matching thread activity arrives
+- `wait-reply`: block until a reply arrives for a specific thread
+
+## Notes
+
+- `body-file` and artifact-heavy flows are supported; prefer them over long inline strings when sending reports or patches.
+- `fetch --unread` and `show --mark-read` depend on per-agent read cursors.
+- If the bundled binary cannot execute on the current host, stop and report the compatibility issue instead of guessing a replacement path or workflow.

skills/inbox/agents/openai.yaml

@@ -0,0 +1,7 @@
+interface:
+  display_name: "Inbox CLI"
+  short_description: "Durable agent coordination CLI"
+  default_prompt: "Use $inbox to coordinate work through the bundled inbox CLI and a SQLite inbox database."
+
+policy:
+  allow_implicit_invocation: true