From claude-recap
Manages .ignore rules with glob patterns to skip archiving matching topic slugs. Add, remove, or list global/project rules via /ignore-topic CLI.
How this skill is triggered — by the user, by Claude, or both
Slash command
/claude-recap:ignore-topicThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Manage `.ignore` rules — glob patterns that prevent matching topic slugs from being archived. Topics still get tagged but archival is skipped (no LLM summary call, no file written).
Manage .ignore rules — glob patterns that prevent matching topic slugs from being archived. Topics still get tagged but archival is skipped (no LLM summary call, no file written).
Parse the user's message for operation (add/remove/list) and patterns. If clear, skip to Step 2.
Otherwise, ask using AskUserQuestion:
What would you like to do?
- Add: Add ignore patterns (topics matching these will not be archived)
- Remove: Remove an existing ignore pattern
- List: Show all current ignore rules
add:
*, ?, [...]).Which scope?
- Global: applies to all projects (~/.memory/.ignore)
- Project: applies only to this project
bash "<plugin_scripts_path>/ignore-topic.sh" add "<scope>" <pattern1> [pattern2 ...]
<scope>: global or the project dir path from SessionStart context.remove:
bash "<plugin_scripts_path>/ignore-topic.sh" remove "<scope>" "<pattern>"
list:
bash "<plugin_scripts_path>/ignore-topic.sh" list "<project_dir>"
<plugin_scripts_path>: from SessionStart injection Plugin scripts path: ...<project_dir>: from SessionStart injection Your persistent memory is stored at <project_dir>| User says | Operation | Scope question? |
|---|---|---|
/ignore-topic add git-* lint-fix | add git-* and lint-fix | Yes |
/ignore-topic | Ask all | Yes |
/ignore-topic list | list | No |
| "don't archive run-tests topics" | add run-tests | Yes |
/ignore-topic remove git-* | remove git-* | Yes |
npx claudepluginhub hatawong/claude-recap --plugin claude-recapPersists user prohibitions as durable rules in AGENTS.md or CLAUDE.md. Apply when the user says 'stop doing X', 'ban this', or 'never do that again'.
Promotes patterns from MEMORY.md to CLAUDE.md or .claude/rules/ for permanent enforcement via /si:promote. Supports auto-detection, scoped targets, and rule distillation.
Guides writing Hookify rules with YAML frontmatter: event types (bash/file/stop/prompt), regex patterns, conditions, and file organization.