agent-environment-retr…
Use when a completed session needs an agent-environment retrospective. Not for an engineering retrospective from telemetry: use engineering-retrospective.
Use when asked to push an already-working artifact to finished composition, texture, and feel. Not for remote or irreversible changes. Not for hygiene routing: use clean-and-true.
$ npx -y skills add OutlineDriven/odin-claude-plugin --skill final-grain --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/final-grainContext preview
The summary Claude sees to decide when to auto-load this skill.
Use when asked to push an already-working artifact to finished composition, texture, and feel. Not for remote or irreversible changes. Not for hygiene routing: use clean-and-true.
name: final-grain description: 'Use when asked to push an already-working artifact to finished composition, texture, and feel. Not for remote or irreversible changes. Not for hygiene routing: use clean-and-true.' disable-model-invocation: true
| Field | Bound contract | |---|---| | Trigger | User wants to push an already working artifact to finished composition, texture, and feel. Not for hygiene routing: use clean-and-true. | | Authority | Reversible local: writes only the named artifact; rollback is version control. No remote mutation. If the artifact is not VCS-tracked, require a human-supplied recovery source before mutation. | | Side effect | Produce a polished local artifact at the named path. No VCS, remote, credential, or paid mutation. | | Done | The artifact exists at the named path with visible polish applied and original functionality confirmed intact. |
1. Confirm the artifact path and that it is working as delivered. Done when: the artifact path is confirmed and the artifact is working. 2. Classify the artifact type (document, code file, config, spec, report, or other). Done when: the artifact type is classified. 3. Propose three polish dimensions (structural coherence, compositional refinement, and presentational polish) unless the human named them. Done when: three polish dimensions are proposed or the human's named targets are adopted. 4. Obtain explicit human confirmation of the polish targets before mutating any file. Done when: the human confirms the polish targets. 5. If the artifact is tracked in version control, note its current committed version as the rollback source. If it is not tracked, require the human to supply an existing recovery source before mutation; otherwise stop as blocked. Done when: the rollback source is noted or the run is blocked for lacking one. 6. Make a copy of the artifact and apply polish to the copy along the confirmed dimensions only. Do not change implementation correctness. Done when: polish is applied to a copy along the confirmed dimensions with implementation correctness unchanged. 7. Verify the polished copy preserves the original functionality: run a smoke test if applicable. Done when: original functionality is confirmed intact. 8. Replace the original artifact with the polished copy. Done when: the polished copy replaces the original at the named path.
A polished artifact at the named path plus a one-paragraph summary of which dimensions were refined and which choices were made in each dimension.
Formerly the ODIN Claude Plugin. The repository URL is unchanged. Outline-Driven Development, nicknamed ODIN, is a highly opinionated code-agent skill library: principles-first engineering, surgical editing, and workflow automation, published as installable
Repo: OutlineDriven/odin-claude-plugin
Use when a completed session needs an agent-environment retrospective. Not for an engineering retrospective from telemetry: use engineering-retrospective.
Use when a repo needs agent setup, AGENTS.md added or made lean, CLAUDE.md audited, or agent instructions scored or pruned. Not for remote, credential,…
Use when a human explicitly asks for a full repository agent-compatibility pass returning a scored report with prioritized fixes. Not for tasks that require…
Use when setting up a project, auditing agent command permissions, or asking which read-only bash commands and domains to allow. Not for remote, credential,…
Use when asked to build or review a CLI intended for coding agents and return flag-driven, pipeline-safe, idempotent design advice. Not for running or…
Use when the user asks to make the skills framework work in a new harness, IDE, or CLI. Not for remote, credential, publish, deploy, or irreversible changes.