From canon
This skill should be used when the user says "log this decision", "remember this choice", "save this to MEMORY.md", "session end", "wrapping up", "let's stop here", or any phrase that signals a significant decision or the close of a working session that should be persisted to MEMORY.md at the project root.
How this skill is triggered — by the user, by Claude, or both
Slash command
/canon:decision-logThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Append structured entries to `MEMORY.md` at the project root so that future sessions can read decisions and session summaries instead of re-deriving them.
Append structured entries to MEMORY.md at the project root so that future sessions can read decisions and session summaries instead of re-deriving them.
This skill handles two kinds of writes: decision entries (a single significant choice the user wants persisted) and session summaries (everything accomplished in the current session).
Decision entries — trigger on phrases like:
Session summaries — trigger on phrases like:
If the user's phrasing is ambiguous, ask which kind of write they want before proceeding.
Look for MEMORY.md at the project root (the same directory as package.json, pyproject.toml, .git, or wherever the user has been working). If it doesn't exist, suggest running the canon-init skill first and ask whether to create a stub MEMORY.md now or skip the write.
For decision entries, format as:
## YYYY-MM-DD: <one-line decision title>
- **What was decided:** <the choice>
- **Why:** <the reason>
- **What was rejected and why:** <alternatives considered and why they lost>
- **Files / areas affected:** <if applicable>
For session summaries, format as:
## YYYY-MM-DD: Session summary
- **Worked on:** <high-level summary>
- **Completed:** <what shipped>
- **In progress:** <what's left>
- **Decisions made:** <link to decision entries above or inline list>
- **Next session priorities:** <what to pick up next>
Pull session content from the conversation history rather than asking the user to repeat themselves. Confirm the draft with the user before writing.
Use Edit or append-mode Write to add the new entry to the top of the section, just below the header, so the most recent entries are visible first. Never delete or modify existing entries.
Show the user the appended block and the line range in MEMORY.md. Offer to also update ERRORS.md if the session involved a failure pattern worth logging there.
npx claudepluginhub orthogon-ai-labs/canonGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.