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 replacing a greenfield attempt with a clean pad of verified requirements. Not for rebuilding from primitives: use from-first-principle. Requires a committed, VCS-recoverable attempt.
$ npx -y skills add OutlineDriven/odin-claude-plugin --skill fromzero --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/fromzeroContext preview
The summary Claude sees to decide when to auto-load this skill.
Use when replacing a greenfield attempt with a clean pad of verified requirements. Not for rebuilding from primitives: use from-first-principle. Requires a committed, VCS-recoverable attempt.
name: fromzero description: 'Use when replacing a greenfield attempt with a clean pad of verified requirements. Not for rebuilding from primitives: use from-first-principle. Requires a committed, VCS-recoverable attempt.' disable-model-invocation: true
| Field | Bound contract | |---|---| | Trigger | The user wants to replace a greenfield attempt with a clean pad carrying only proven requirements. | | Authority | Human-gated: requires explicit human invocation before acting; otherwise reversible local: writes only VCS-tracked targets (show the exact set before mutation); rollback is version control. No remote mutation. | | Side effect | A new greenfield attempt on a clean pad; the old attempt remains recoverable through VCS. | | Done | A new greenfield attempt exists carrying only verified requirement statements, and the old attempt is reachable in VCS history. |
1. Verify the current attempt is committed and recoverable through version control. If any uncommitted or untracked work exists, stop and require the human to commit or discard it first. Done when: the working tree is clean and the attempt is in VCS history.
2. Filter requirements to verified only. Carry only requirements that have been verified; discard unproven, speculative, or accumulated scope. Done when: the proven-requirements set is listed and each entry is verified.
3. Branch from the recoverable commit to clear the working surface. Before any mutation, list the exact set of VCS-tracked targets the clearance will delete (the tracked files of the old attempt) and show that set; the mutation is a branch creation followed by the clearance of exactly those listed targets. The old attempt survives in VCS history while the working surface is cleared on the new branch. Do not use an unspecified reset alternative. Done when: the exact target set was shown before mutating, the working surface is clear on the new branch, and the old attempt is reachable in VCS.
4. Write the requirement list into the new pad. Write only the proven requirements as statements. Do not reintroduce discarded scope. Done when: the new pad contains only verified requirement statements.
A new greenfield attempt on a clean pad carrying only verified requirement statements, with the old attempt recoverable through VCS history.
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.