adr-log
Skill0 starsChanged 4 months agoMITView on GitHub
Record significant architecture or execution decisions so their reasoning survives the session. Trigger on requests like record this decision, write an ADR, capture the trade-off, or supersede the old decision.
In this repository37 other files it gives its agents.
- adversarial-reviewSkill
- ai-writing-detoxSkill
- autonomous-video-pipelineSkill
- browser-tutorial-videoSkill
- codex-execution-packSkill
- continuity-ledgerSkill
- github-mobile-opsSkill
- github-seoSkill
- idea-validationSkill
- illustration-promptSkill
- issue-control-loopSkill
- justdoitSkill
- lazy-seniorSkill
- long-context-dispatchSkill
- openclaw-councilSkill
- parallel-worktreesSkill
- playwrightSkill
- product-councilSkill
- product-shapingSkill
- public-artifact-laneSkill
- readme-generatorSkill
- screenshotSkill
- session-to-postSkill
- spec-bundleSkill
- spreadsheetSkill
- spreadsheet-tabular-analysisSkill
- spreadsheet-xlsx-editSkill
- tool-scoutSkill
- transcribe-basicSkill
- transcribe-diarizeSkill
- transcribeSkill
- triage-findingSkill
- video-builderSkill
- video-render-runSkill
- video-spec-builderSkill
- visual-explainerSkill
- work-shapingSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 10
Do not use when: the choice is a tiny implementation detail or the decision is still too fuzzy to commit.
- Distinctive: line 24
If the rationale was reconstructed after the fact, label it as reconstructed instead of pretending it was captured live.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
ADR Log
Metadata
- Trigger when: the task is to capture one real stack, schema, workflow, boundary, or vendor decision with durable rationale.
- Do not use when: the choice is a tiny implementation detail or the decision is still too fuzzy to commit.
Skill Purpose
Capture one expensive-to-rediscover decision so the chosen path, rejected alternatives, and revisit conditions survive beyond the current chat session.
Instructions
- Gate the decision first. Confirm what was chosen, what was rejected, and why this is worth preserving. If the choice is still fuzzy, route to
$product-councilor$spec-bundleinstead. Read~/.codex/skills/adr-log/references/storage-rules.mdonly when you need the storage surface and~/.codex/skills/adr-log/references/entry-template.mdonly when you need the exact entry shape. - Record one entry in the right durable place: repo ADR file when the decision belongs to code, GitHub issue summary when issue state is canonical, or both when the project already uses
$issue-control-loop. - Validate the record. Make sure it names context, options, chosen path, trade-offs, consequences, revisit conditions, and whether this entry supersedes an older decision.
Non-Negotiable Acceptance Criteria
- One entry covers one real decision.
- Rejected alternatives and trade-offs are explicit; they are not optional filler.
- If the rationale was reconstructed after the fact, label it as reconstructed instead of pretending it was captured live.
- Superseded decisions are marked as superseded, not silently overwritten.
Output
- The written ADR or issue summary saved to its durable destination.
- A short summary of the chosen path, the main trade-off, and the revisit trigger.
- The destination path or issue location so the decision can be reopened later.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

