| name | ccr-recall |
| user-invocable | true |
| description | Use when the user asks to recall, search, or continue past conversations. Triggers on "what did we discuss", "continue where we left off", "remember when", "as I mentioned", "you suggested", "we decided", "search my conversations", "find the conversation where", "what did we work on". Also triggers on implicit signals like past-tense references ("the bug we fixed"), possessives without context ("my project"), or assumptive questions ("do you remember").
|
Tools
Three commands retrieve data. For full option catalogs, load references/tool-reference.md.
Under a plugin install this skill is invoked as /ccrecall:ccr-recall; under a vendored install, /ccr-recall.
Semantic recall is active: ccrecall search fuses keyword ranking with chunk-vector similarity by default.
recent_chats.py — retrieve recent sessions with full transcript:
ccrecall recent --n 3
search_conversations.py / Entrypoint A — search sessions, returns ranked session cards (compact, no transcript):
ccrecall search --query "keyword"
Each card carries a relevance score (normalized 0–1; null for a single result or the unranked LIKE path), project, git_branch, date, topic, exchange/file/commit counts, and a handle. The JSON envelope includes a ranked field (false on the LIKE-only fallback). To read the full session, use ccrecall tail <handle>.
search_conversations.py / Entrypoint B — search matched exchanges, returns bounded snippets:
ccrecall search-messages --query "keyword"
Each snippet carries a (handle, exchange_index, timestamp) locator plus bounded user/assistant excerpts. Semantically ranked by chunk distance; exits 0 with an empty result when the vector index is unavailable. No keyword fallback in this release.
session_tail.py — full-fetch drill-in for any handle from A or B:
ccrecall tail <handle>
Workflow
- Identify the lens from user intent:
| User Says | Lens |
|---|
| "where were we", "recap" | restore-context |
| "gaps", "struggling" | find-gaps |
| "mentor", "review process" | review-process |
| "retro", "project review" | run-retro |
| "decisions", "CLAUDE.md" | extract-decisions |
| "bad habits", "antipatterns" | find-antipatterns |
Load references/lenses.md for per-lens parameters, core questions, and supplementary search patterns.
-
Gather context using lens-appropriate tools:
- For recent context:
ccrecall recent --n N
- For session discovery:
ccrecall search --query "keywords" — returns scored cards; triage by score and topic before tailing
- For a specific exchange:
ccrecall search-messages --query "phrase" — returns bounded snippets with locators
- To open a full session:
ccrecall tail <handle> (drill-in after A or B)
-
Apply lens questions to analyze the retrieved conversations.
-
Deepen the search if initial results are insufficient:
- Retrieve more sessions:
--n 20 on recent, or --max-results 10 on search
- Search for specific terms that surfaced
- Filter by project:
--project projectname
- Filter by session:
--session <uuid-prefix> (when a specific session ID is known)
- If 2 rounds of deepening yield no new relevant sessions, synthesize from available data.
Query Construction
Search terms should be content-bearing words that discriminate between sessions — high information value words that are rare enough to rank relevant sessions above irrelevant ones. BM25 ranking (when FTS5 is available) weights rare terms higher automatically.
Include: specific nouns, technologies, concepts, project names, domain terms, unique phrases. More terms improve ranking precision.
Exclude: generic verbs ("discuss", "talk"), time markers ("yesterday"), vague nouns ("thing", "stuff"), meta-conversation words ("conversation", "chat") — these appear in nearly every session and add noise rather than signal.
Algorithm:
- Extract substantive keywords from user request
- If 0 keywords, ask for clarification ("Which project specifically?")
- If 1+ specific terms, search with those terms; use
--project to narrow scope
Synthesis
Principles
- 3-5 key findings, not exhaustive lists — each specific (file paths, dates, project names) and backed by a quote or reference.
- Make it actionable — every finding suggests a response.
Structure
## [Analysis Type]: [Scope]
### Summary
[2-3 sentences]
### Findings
[Organized by whatever fits: categories, timeline, severity]
### Patterns
[Cross-cutting observations]
### Recommendations
[Actionable next steps]
Length
Default: 300-500 words. Expand only when data warrants it.
