recall（recall 搜索）

Recall

Usage — /recall

Load branch-scoped context from .context/ and optionally browse project-scoped memory in .rulesync/.

STEP 1: Load Branch Memory

Determine Context File Path and Load Content

Run the co-located script to resolve paths and load any existing context. Use the skill base directory from the header above:

bash "<skill-base-directory>/context-path.sh" --list-rules

Parse the structured output:

• BRANCH= — current branch name
• SLUG= — filename slug
• CONTEXT_FILE= — full path to context file
• STATUS=found|not_found — whether context exists
• Content after ---CONTENT--- — existing context file contents (if found)
• Content after ---RULES--- — list of project rule files (if present)

Present Branch Context

If found, present its contents:

## Branch Context Loaded: {branch}

**Intent**: {summary of intent section}

**Key Patterns**: {count} patterns documented

**Known Gaps**: {count} gaps tracked

**References**: {count} references available

---

{Full context file contents}

If not found, inform the user:

No saved context found for branch {branch}. Run /remember and choose Branch to create context for this branch.

STEP 2: Project Memory Summary (optional)

After presenting branch context, offer to show project memory:

Project rules and learnings in .rulesync/rules/ load automatically based on file-pattern globs. Want to browse the project memory files?

If the user says yes:

1. Present the rules listing from the ---RULES--- section of the script output
2. User can request to read specific memory files for details
3. This is informational — project rules auto-load during normal work via globs

Notes

• Context files are stored in the main repo root, shared across all worktrees
• Context persists even when worktrees are deleted
• Branch context files are gitignored — this is personal/local context, not shared
• Project memory (.rulesync/rules/) is committed and shared with the team