Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/cfircoo/claude-code-toolkit/install-toolkitgit clone --depth 1 https://github.com/cfircoo/claude-code-toolkitWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00009 | $0.01580 |
| Opus 5 | $0.00005 | $0.00790 |
| Sonnet 5 | $0.00002 | $0.00316 |
| Haiku 4.5 | $0.00001 | $0.00158 |
Grade F, and why
install-toolkit scanned grade F with 4 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Downloads and executes remote codehighSupply chain
curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.
- If UV not found, offer to install: `curl -LsSf https://astral.sh/uv/install.sh | sh` Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
cp ~/.claude/settings.json ~/.claude/settings.json.bak 2>/dev/null Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
- **Damage Control**: Security hooks that block dangerous commands (rm -rf, git reset --hard, etc.) and protect sensitive files (.env, ~/.ssh/, credentials) Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- If UV not found, offer to install: `curl -LsSf https://astral.sh/uv/install.sh | sh` How it starts
The opening of the file, as written. The whole thing — 150 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Available components:
- Skills (12): git, pytest-best-practices, sqlalchemy-postgres, debug-like-expert, create-plans, spec-interview, manage-subagents, manage-hooks, manage-slash-commands, manage-skills, create-meta-prompts, damage-control
- Agents (7): git-ops, db-expert, pytest-writer, fullstack-manager, fullstack-api-specialist, fullstack-ui-designer, fullstack-qa-debugger
- Commands (6): commit, push, pr, ship, db, spec-interview
- Hooks: hooks.json + hook scripts
- Damage Control: Security hooks that block dangerous commands (rm -rf, git reset --hard, etc.) and protect sensitive files (.env, ~/.ssh/, credentials)
- Extras: statusline.sh, settings.json
-
Check dependencies: Run bash commands to check for jq, git, and platform-specific tools (tac/gtac)
- Report status to user
- If missing, ask if they want to install them
-
Check Claude Code CLI:
- If installed: show version, offer to check for updates
- If not installed, offer to install:
- macOS:
brew install --cask claude-code(preferred) - Linux:
npm install -g @anthropic-ai/claude-code@latest
- macOS:
- If package manager not available: show manual install instructions
-
Ask installation mode using AskUserQuestion:
- "Install All" - Copy everything (recommended)
- "Select by Category" - Choose which folders to install
- "Custom Selection" - Pick individual items
-
If "Select by Category", ask for each category:
- Skills: All / Skip
- Agents: All / Skip
- Commands: All / Skip
- Hooks: All / Skip
- Damage Control: Yes / No (security hooks)
- Statusline: Yes / No
- Settings: Merge / Skip
-
If "Custom Selection", for each category ask which specific items to install
-
Execute installation:
# Create directories mkdir -p ~/.claude/{skills,agents,commands,hooks} # Copy selected skills (replace existing) rm -rf ~/.claude/skills/SKILL_NAME && cp -r TOOLKIT/skills/SKILL_NAME ~/.claude/skills/ # Copy selected agents cp TOOLKIT/agents/AGENT.md ~/.claude/agents/ # Copy selected commands cp TOOLKIT/commands/COMMAND.md ~/.claude/commands/ # Copy hooks (backup existing) cp ~/.claude/hooks.json ~/.claude/hooks.json.bak 2>/dev/null cp TOOLKIT/hooks.json ~/.claude/hooks.json # Copy statusline cp TOOLKIT/statusline.sh ~/.claude/statusline.sh && chmod +x ~/.claude/statusline.sh # Merge settings (if jq available) cp ~/.claude/settings.json ~/.claude/settings.json.bak 2>/dev/null jq -s '.[1] * .[0]' TOOLKIT/settings.json ~/.claude/settings.json > /tmp/merged.json && mv /tmp/merged.json ~/.claude/settings.json -
Install Damage Control (if selected):
- Check for UV runtime:
which uv - If UV not found, offer to install:
curl -LsSf https://astral.sh/uv/install.sh | sh - Create hooks directory:
mkdir -p ~/.claude/hooks/damage-control - Copy Python scripts:
cp TOOLKIT/skills/damage-control/scripts/*.py ~/.claude/hooks/damage-control/ - Copy patterns.yaml:
cp TOOLKIT/skills/damage-control/scripts/patterns.yaml ~/.claude/hooks/damage-control/ - Make executable:
chmod +x ~/.claude/hooks/damage-control/*.py - Merge settings with damage-control hooks and permissions:
# Backup existing cp ~/.claude/settings.json ~/.claude/settings.json.bak 2>/dev/null # Merge hooks and permissions jq -s ' def merge_hooks: (.[0].hooks.PreToolUse // []) + (.[1].hooks.PreToolUse // []) | unique_by(.matcher); def merge_permissions: { deny: ((.[0].permissions.deny // []) + (.[1].permissions.deny // []) | unique), ask: ((.[0].permissions.ask // []) + (.[1].permissions.ask // []) | unique) }; .[0] * .[1] | .hooks.PreToolUse = (.[0] | merge_hooks) | .permissions = ([.[0], .[1]] | merge_permissions) ' ~/.claude/settings.json TOOLKIT/skills/damage-control/scripts/settings-template.json > /tmp/merged.json && mv /tmp/merged.json ~/.claude/settings.json - Verify files exist:
~/.claude/hooks/damage-control/bash-tool-damage-control.py~/.claude/hooks/damage-control/edit-tool-damage-control.py~/.claude/hooks/damage-control/write-tool-damage-control.py~/.claude/hooks/damage-control/patterns.yaml
- IMPORTANT: Remind user to restart Claude Code for hooks to take effect
- Tell user to run
/hooksafter restart to verify registration
- Check for UV runtime:
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 150 lines · 9 tokens per session scan F adb2117f428e
install-toolkit is a command published in the GitHub repository cfircoo/claude-code-toolkit (17 stars, last pushed 5mo ago), licensed MIT. It adds 9 tokens to every session and 1,580 once invoked, about $0.0000 per session on Opus 5. A static security scan graded it F with 4 findings (downloads and executes remote code, reads agent configuration directories, recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
git
Git operations with intelligent commit messages and workflow optimization.
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.