Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/thunter009-colgrep
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/colgrep
Tools it pre-approves 1
From allowed-tools in SKILL.md: the agent may use these without asking you each time.
Bash(colgrep:*)
What's inside
- Skill name
colgrep- Description
- Search code by meaning with colgrep instead of literal grep.
- License
- MIT
- Allowed tools
- Bash(colgrep:*)
- Instructions
- 80 lines in SKILL.md · view on GitHub
Before installing
- Pre-approves some shell commands
Files (1)
- SKILL.md
SKILL.md (first lines)
# Semantic Code Search
`colgrep` is installed at `~/.cargo/bin/colgrep` — a semantic code search CLI using ColBERT embeddings.
**Use `colgrep` as your PRIMARY search tool** instead of Grep/Glob for code search.
## Quick Reference
```bash
# Basic semantic search
colgrep "<natural language query>" --results 10 # Basic search
colgrep "<query>" -k 25 # Exploration (more results)
colgrep "<query>" ./src/parser # Search in specific folder
colgrep "<query>" ./src/main.rs # Search in specific file
colgrep "<query>" ./crate-a ./crate-b # Search multiple directories
# File filtering
colgrep --include="*.rs" "<query>" # Include only .rs files
colgrep --include="*.{rs,md,py}" "<query>" # Multiple file types
colgrep --exclude="*.test.ts" "<query>" # Exclude test files
colgrep --exclude-dir=vendor "<query>" # Exclude vendor directory
# Pattern-only search (no semantic query needed)
colgrep -e "<pattern>" # Search by pattern only
colgrep -e "async fn" --include="*.rs" # Pattern search with file filter
# Hybrid search (text + semantic)
colgrep -e "<text>" "<semantic query>" # Hybrid: text + semantic
colgrep -e "<regex>" -E "<semantic query>" # Hybrid with extended regex
colgrep -e "<literal>" -F "<semantic query>" # Hybrid with fixed string
colgrep -e "<word>" -w "<semantic query>" # Hybrid with whole word match
# Output options
colgrep -l "<query>" # List files only
colgrep -n 6 "<query>" # Show N context lines
colgrep -c "<query>" # Show full function/class content
colgrep --json "<query>" # JSON output for scripting
# Index management
colgrep init /path/to/project # Build/update index
colgrep init -y # Auto-confirm large projects
colgrep status # Check index health
colgrep clear # Clear current project index
```
## When to Use What
| Task | Tool |
| ------------------------------- | -------------------------------------------- |
| Find code by intent/description | `colgrep "query" -k 10` |
| Explore/understand a system | `colgrep "query" -k 25` |
| Search by pattern only | `colgrep -e "pattern"` |
| Know text exists, need context | `colgrep -e "text" "semantic query"` |
| Search specific file type | `colgrep --include="*.ext" "query"` |
| Search multiple directories | `colgrep "query" ./src ./lib ./api` |
| Exact string/regex match only | Built-in `Grep` tool |
| Find files by name | Built-in `Glob` tool |
## Key Rules
1. **Default to `colgrep`** for any code search task
2. **Auto-indexes on first search** — no need to `colgrep init` manually (but init is faster for large repos)
3. **Increase `-k`** when exploring (20-30 results)
4. **Use `-e`** for hybrid text+semantic filtering
5. **Use `--exclude-dir`** to filter noise (tests, vendors, generated code)
6. **Subagents should also use `colgrep`** — when spawning Task/Explore agents, instruct them to use colgrep
7. **Fall back to Grep** only for exact literal/regex matches where semantic search adds no value
## Why Semantic Search Matters
Keyword grep misses code that implements a concept without using the expected words. Example:
**Task:** "Find all authentication logic"
- `grep -rl "auth\|jwt\|login\|password"` finds files with those literal strings
- But misses: TOTP/2FA in `crypto_utils.py`, role-based access in `decorators.py` (`require_role`), RBAC policies in `policies.py` (`has_permission`, `enforce_authorization`)
- `colgrep "authentication and authorization logic" -k 15` finds ALL of them because it understands that TOTP, role guards, and permission checks are auth concepts
When searching for a concept (not a literal string), always use colgrep.From thunter009/skills, MIT.
More from thunter009/skills
- 0.570score
decision-canvas
thunter009
Present batched user decisions as an interactive atrium canvas with inline grounding — code snippets with real line numbers, click-to-open file links, and…
SkillAFreshness: AClaude Code +9 - 0.529score
clean
thunter009
End-of-work cleanup sweep for a multi-agent repo — safely land genuine unpushed commits, squash open PRs, prune merged branches and worktrees, clean the…
SkillBFreshness: BClaude Code +9 - 0.526score
autoreview
thunter009
Run a structured code review (Codex default, Claude optional) as a closeout check on a local or PR branch before commit or ship.
SkillBFreshness: BClaude Code +9 - 0.520score
bug-hunt
thunter009
Run a multi-phase bug hunt with Codex, Gemini, or Claude.
SkillBFreshness: BClaude Code +9 - 0.518score
create-pull-request
thunter009
Open a GitHub pull request with repo-specific PR conventions.
SkillBFreshness: BClaude Code +9 - 0.514score
agents-md
thunter009
Create repo-specific AGENTS.md guardrails from the codebase.
SkillBFreshness: BClaude Code +9 - 0.479score
enrich-ticket-detail
thunter009
Flesh out thin or empty beads/Todoist items by mining git history, code-at-HEAD, closed tickets, and the session journal, then verify every drafted claim…
SkillCFreshness: CClaude Code +9 - 0.476score
codebase-knowledge-graph
thunter009
Build an interactive knowledge graph of any codebase via the Understand-Anything pipeline, with a batch-integrity guard the raw plugin lacks.
SkillCFreshness: CClaude Code +9 - 0.476score
docs-organize
thunter009
Reorganize loose files already in a folder into its subfolder taxonomy, with hash-based dedup, dry-run preview, and collision-safe moves.
SkillCFreshness: CClaude Code +9 - 0.474score
devbox-cleanup
thunter009
Clean up a project checkout on a remote dev box — tear down a stale agent swarm, verify unpushed work is superseded, reset the checkout to origin, and restart…
SkillCFreshness: CClaude Code +9 - 0.470score
domain-modeling
thunter009
Build and sharpen a project's ubiquitous language — a CONTEXT.md glossary plus sparse ADRs — challenging terms against the glossary and cross-referencing code…
SkillCFreshness: CClaude Code +9 - 0.469score
beads-linear-sync
thunter009
Sync beads and Linear so agent state and human tracker stay aligned.
SkillCFreshness: CClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Jul 14, 2026 (3 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/thunter009/skills
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from thunter009
- 0.570score
decision-canvas
thunter009
Present batched user decisions as an interactive atrium canvas with inline grounding — code snippets with real line numbers, click-to-open file links, and…
SkillAFreshness: AClaude Code +9 - 0.529score
clean
thunter009
End-of-work cleanup sweep for a multi-agent repo — safely land genuine unpushed commits, squash open PRs, prune merged branches and worktrees, clean the…
SkillBFreshness: BClaude Code +9 - 0.526score
autoreview
thunter009
Run a structured code review (Codex default, Claude optional) as a closeout check on a local or PR branch before commit or ship.
SkillBFreshness: BClaude Code +9 - 0.520score
bug-hunt
thunter009
Run a multi-phase bug hunt with Codex, Gemini, or Claude.
SkillBFreshness: BClaude Code +9 - 0.518score
create-pull-request
thunter009
Open a GitHub pull request with repo-specific PR conventions.
SkillBFreshness: BClaude Code +9 - 0.514score
agents-md
thunter009
Create repo-specific AGENTS.md guardrails from the codebase.
SkillBFreshness: BClaude Code +9
Cite this page
colgrep skill. Agent Mart. https://agentmart.dev/skills/thunter009-colgrep (last updated 2026-10-24)