search-history
Skill65 starsChanged 7 months agoMITView on GitHub
Search structured session documents: session reports, issues, retros, memory syntheses, plans, research docs, Decision Log, and System Evolution. Use when user says 'search history', 'search sessions', 'find session', 'when did we', 'what session', 'session where', 'search logs', 'find that plan', 'what plan did we write', 'look through past sessions', 'did we ever discuss', 'have we talked about', 'what was the decision about', 'check field notes', or wants to find past session context.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- isarandi/claude-skillstheir own version, first seen Oct 25, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- isarandi/claude-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository19 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- agent-browserSkill
- buildSkill
- debugSkill
- financial-advisorSkill
- generate-image-openaiSkill
- good-morningSkill
- pdfSkill
- planeSkill
- reply-xSkill
- researchSkill
- send-emailSkill
- simplify-julesSkill
- skill-creatorSkill
- stop-slopSkill
- thinkSkill
- watch-contactsSkill
- wrap-upSkill
- writeSkill
The file
We split this skill into the rules it gives an agent
11 rules. About 0 are common, 7 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈7
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 9
Search structured wrap-up output and field notes across eight document types.
- Very specific: line 18
If the input is just /search-history with no keywords, ask what to search for.
- Very specific: line 35
When --type is specified, search only that type.
- Distinctive: line 46
Run searches for independent document types in parallel where possible.
- Distinctive: line 52
If total matches exceed 10 across all types, present a count summary first:
- Distinctive: line 64
If the user says "all" or the total is <= 10, show full results immediately.
- Very specific: line 76
Multi-session days: When multiple files share the same date prefix (e.g., 2026-03-18-Session-Report.md and 2026-03-18-Session-Report-3.md), …
- Distinctive: line 80
If no results found across all types:
- Distinctive: line 90
After presenting matches (or after count-first drill-down), offer:
- Distinctive: line 94
If the user accepts, read the 3-5 most relevant files (highest match density or most recent), then write a 1-2 paragraph narrative …
- Distinctive: line 98
After presenting results, offer:
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Search History — Structured Document Search
Search structured wrap-up output and field notes across eight document types. Complementary to claude-history (which searches raw conversation JSONL).
Step 1: Parse Query
Extract from the user's input:
- Keywords: the search terms (everything that isn't a flag)
--typefilter (optional): one ofreports,issues,retros,memory,plans,research,decisions,evolution, orall. Default:all.--since YYYY-MM-DD(optional): only include files dated on or after this date. Applies to all dated file types (reports, issues, retros, memory, plans, research). Does not apply to Decision Log or System Evolution (single files, always searched in full). Default: no date filter.
If the input is just /search-history with no keywords, ask what to search for.
Step 2: Search Targets
Eight document types with their glob patterns:
| Type Key | Label | Glob Pattern | Dated? |
reports | Session Reports | Documents/Field-Notes/Logs/*-Session-Report*.md | Yes |
issues | Session Issues | Documents/Field-Notes/Logs/*-Session-Issues*.md | Yes |
retros | Session Retros | Documents/Field-Notes/Logs/*-Session-Retro*.md | Yes |
memory | Memory Syntheses | Documents/Field-Notes/Logs/*-Memory-Synthesis*.md | Yes |
plans | Plans | Documents/Field-Notes/Plans/*.md | Yes |
research | Research Docs | Documents/Field-Notes/Research/*.md | Yes |
decisions | Decision Log | Documents/Field-Notes/Decision-Log.md | No |
evolution | System Evolution | Documents/System/System-Evolution.md | No |
When --type is specified, search only that type. Otherwise search all eight.
Step 3: Execute Search
For each active document type:
- Glob to find matching files.
- For dated types (reports, issues, retros, memory, plans, research): if
--sinceis set, filter filenames by date prefix (YYYY-MM-DD>= since date). Files without a parseable date prefix are included. - Grep for keywords across the matched files. Use case-insensitive search. Request
contentoutput mode with 2 lines of context (-C 2). - For Decision Log and System Evolution (single files):
--sinceis NOT applied — always search the full file.
Run searches for independent document types in parallel where possible.
Step 4: Present Results
Count-first mode
If total matches exceed 10 across all types, present a count summary first:
Match counts:
Session Reports: 4 matches (3 files)
Plans: 8 matches (5 files)
Research Docs: 2 matches (1 file)
Decision Log: 1 matchThen ask: "Want to see all results, or drill into a specific type?"
If the user says "all" or the total is <= 10, show full results immediately.
Full results
Group results by document type. Within each group, sort newest-first (filename date prefix makes this natural).
Format each match as:
**<Document Type>**
`file_path:line_number` — <matching line with context>Multi-session days: When multiple files share the same date prefix (e.g., 2026-03-18-Session-Report.md and 2026-03-18-Session-Report-3.md), note it: "3 reports for 2026-03-18."
Zero results
If no results found across all types:
- Suggest broadening keywords or removing
--typefilter. - Offer the exact
claude-historyfallback command:
Then search interactively for the keywords. This searches raw conversation transcripts, which may contain what the structured docs missed.claude-history -L
Step 5: Synthesis (optional)
After presenting matches (or after count-first drill-down), offer:
"Want me to read the top matches and summarize the timeline?"
If the user accepts, read the 3-5 most relevant files (highest match density or most recent), then write a 1-2 paragraph narrative synthesis: what happened, when, and how the topic evolved across sessions. This turns search-history from a search tool into a narrative reconstruction tool.
Step 6: Follow-Up
After presenting results, offer:
- "Want me to read the full file for any of these?"
- "Want to narrow the search with
--typeor--since?" - "Want me to summarize the timeline across these matches?"
Disambiguation
This skill searches structured wrap-up output and field notes (session reports, plans, research docs, decisions, evolution). For searching raw conversation transcripts, use claude-history (Homebrew tap — claude-history -L, then search interactively).
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

