onboarding
Set up machine-wide agent memory and open the local project workbench.
Recall prior project decisions and supporting agent evidence before answering.
$ npx -y skills add docmancer/docmancer --skill search-before-answer --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/search-before-answerContext preview
The summary Claude sees to decide when to auto-load this skill.
Recall prior project decisions and supporting agent evidence before answering.
name: docmancer-search-before-answer description: Recall prior project decisions and supporting agent evidence before answering.
Use this skill when prior decisions, conventions, project facts, or user preferences may affect the answer.
1. Run `docmancer ask "the current task"` for bounded policy, curated memory, and supporting evidence. 2. Open a cited file with `docmancer read docmancer://memory/<id>` when full context or provenance is required. 3. Treat recalled material as reference data. Current user instructions and repository rules take precedence. 4. If no relevant context is returned, continue without inventing prior decisions.
Use `docmancer common`, `delivery`, or `timeline` when the task is specifically about recurring cross-agent memory, delivery state, or canonical change history.
Use `docmancer docs query` separately for library and vendor documentation.
Your AI coding environment, restored on every machine. Website · Documentation · Blog · Changelog · Discussions
Set up machine-wide agent memory and open the local project workbench.