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/vv1nn/skill-guide/guidegit clone --depth 1 https://github.com/VV1NN/skill-guideWhat 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 | $0.00067 | $0.03220 |
| Opus 5 | $0.00034 | $0.01610 |
| Sonnet 5 | $0.00013 | $0.00644 |
| Haiku 4.5 | $0.00007 | $0.00322 |
Grade E, and why
guide scanned grade E with 3 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Downloads and executes remote codehighSupply chain
curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.
- forge: `curl -L https://foundry.paradigm.xyz | bash && foundryup` Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
[ -n "$TEMP_DIR" ] && rm -rf "$TEMP_DIR" Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
git curl jq gh docker; do How it starts
The opening of the file, as written. The whole thing — 365 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/guide -- Skill Navigator
You are a skill navigator helping the user understand their installed Claude Code skills and commands.
Detect the user's language from their input and respond entirely in that language.
Step 1: Determine the mode
Based on the argument provided by the user ($ARGUMENTS):
- No arguments or empty --> Run Overview Mode
--check--> Run Dependency Check Mode--diff <repo-url-or-skill-pack-name>--> Run Diff Mode- Argument matches an installed skill or command name (e.g.,
hunt,recon,web3-audit) --> Run Deep Dive Mode - Argument is a sentence or goal description (e.g., "I want to test website security") --> Run Recommendation Mode
Step 2: Scan installed skills and commands
Use the Bash tool to scan the filesystem:
# Use bash explicitly to avoid zsh glob no-match errors
bash -c '
# Scan user-level skills
for f in ~/.claude/skills/*/SKILL.md; do
[ -f "$f" ] || continue
[ -s "$f" ] || continue # skip empty files
name=$(grep "^name:" "$f" | head -1 | sed "s/name: *//")
# Fallback: use directory name if no name field
[ -z "$name" ] && name=$(basename "$(dirname "$f")")
desc=$(grep "^description:" "$f" | head -1 | sed "s/description: *//" | cut -c1-150)
[ -z "$desc" ] && desc="(no description available)"
echo "SKILL|$name|$desc"
done
# Scan project-level skills
for f in .claude/skills/*/SKILL.md; do
[ -f "$f" ] || continue
[ -s "$f" ] || continue
name=$(grep "^name:" "$f" | head -1 | sed "s/name: *//")
[ -z "$name" ] && name=$(basename "$(dirname "$f")")
desc=$(grep "^description:" "$f" | head -1 | sed "s/description: *//" | cut -c1-150)
[ -z "$desc" ] && desc="(no description available)"
echo "SKILL|$name|$desc (project)"
done
# Scan user-level commands
for f in ~/.claude/commands/*.md; do
[ -f "$f" ] || continue
[ -s "$f" ] || continue
cmd=$(basename "$f" .md)
desc=$(grep "^description:" "$f" | head -1 | sed "s/description: *//" | cut -c1-150)
[ -z "$desc" ] && desc="(no description available)"
echo "CMD|/$cmd|$desc"
done
# Scan project-level commands
for f in .claude/commands/*.md; do
[ -f "$f" ] || continue
[ -s "$f" ] || continue
cmd=$(basename "$f" .md)
desc=$(grep "^description:" "$f" | head -1 | sed "s/description: *//" | cut -c1-150)
[ -z "$desc" ] && desc="(no description available)"
echo "CMD|/$cmd|$desc (project)"
done
'
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.
- yesterday First seen · 365 lines · 67 tokens per session scan E bb688ef8d100
guide is a command published in the GitHub repository VV1NN/skill-guide (5 stars, last pushed 4mo ago), licensed MIT. It adds 67 tokens to every session and 3,220 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it E with 3 findings (downloads and executes remote code, recursive force delete, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
quick-commit
Create a conventional commit with an auto-generated message.
review-pr
Review a pull request and provide structured feedback.
edit-searchable
Add plugins to, or remove them from, the zero-token searchable tier.
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.