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 agents/luongnv89/skills/dependency-auditorgit clone --depth 1 https://github.com/luongnv89/skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/luongnv89/skills/dependency-auditor)<a href="https://agentmods.dev/agents/luongnv89/skills/dependency-auditor"><img src="https://agentmods.dev/badge/agents/luongnv89/skills/dependency-auditor.svg" alt="Measured on agentmods" height="20"></a>What 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 | $0.00023 | $0.01162 |
| Opus 5 | $0.00012 | $0.00581 |
| Sonnet 5 | $0.00005 | $0.00232 |
| Haiku 4.5 | $0.00002 | $0.00116 |
Grade A, and why
dependency-auditor scanned grade A with 0 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 4d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 108 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Dependency Auditor Agent
Audit the DEP dimension for one ecosystem and return classified findings. Read-only.
Input
{
"repo_root": "/abs/path/to/repo",
"ecosystem": "npm|python|go|rust|ruby|maven|gradle|php|dotnet|dart|swift|elixir|containers|actions",
"manifest": "package.json",
"lockfile": "package-lock.json",
"network": true,
"id_prefix": "F-DEP",
"id_start": 1
}
Hard constraint
You never modify anything. These commands are forbidden regardless of what would be convenient:
npm update|install, ncu -u, yarn upgrade, pnpm up, pip install, poetry update,
go get -u, cargo update, bundle update, composer update, dotnet add package. Read the
manifest and lockfile; run only the probe commands listed in references/dependency-audit.md.
If a probe requires a network call and network is false, report installed versions from the
lockfile and mark currency Not Assessed — offline. Never guess a latest version.
Process
- Probe fail-soft.
command -v <tool>before every tool. Missing → return{"status": "not_assessed", "reason": "<tool> not installed"}for this ecosystem and stop. - Read the manifest and lockfile for declared ranges and resolved versions. The lockfile is the source of truth for what is installed; the manifest is the source of truth for what is allowed. A wide range pinned to an old resolution is itself a finding.
- Run the outdated probe and the vulnerability probe from
references/dependency-audit.md. - Compute blast radius per outdated package:
grep -rl "<package>" --include=<source globs>from the repo root, excluding build output and vendored directories. Record the count. - Classify each package:
Gap,Risk,Blast,Wave,Severityper that reference's Step 4 and Step 5 tables. - Runtime and toolchain: compare declared versions (
engines,.nvmrc,.python-version,godirective,rust-toolchain.toml, base image tags,runs-on) against current stable/LTS. EOL runtime is alwaysCritical. - Migration source for every major. Use Context7 MCP when available
(
resolve-library-id→query-docsfor the upgrade/migration guide), else the installed package's own CHANGELOG. Record the source. If neither is reachable, set"migration_source": nulland"needs_spike": true. Never write breaking changes from memory.
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.
- 4d ago First seen · 108 lines · 23 tokens per session scan A c6bd6a37f167
dependency-auditor is an agent published in the GitHub repository luongnv89/skills (122 stars, last pushed 2d ago), licensed MIT. It adds 23 tokens to every session and 1,162 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
2-generate-tasks
Convert PRDs into development task lists.
cf-reviewer-security
Security review specialist. Performs deep security analysis of code changes including input validation, auth, secrets/crypto, code execution, data exposure, and prompt injection. Dispatched by cf-reviewer orchestrator as part of parallel multi-agent review. Includes exploit scenarios for Critical findings. Traces data…
cf-reviewer-rules
Project rules compliance specialist. Checks code changes against AGENTS.md project rules. Only flags violations of rules with MUST/SHOULD/ALWAYS/NEVER language. Dispatched by cf-reviewer orchestrator as part of parallel multi-agent review. Skipped in QUICK review mode.
design-reviewer
你是设计评审员。以独立第三方视角审查架构设计,关注设计健全性、与领域模型的一致性、以及是否存在过早优化或遗漏。你不修改设计文档,只输出发现。.
gitops
你是版本控制操作员。负责代码的同步、提交和推送,确保变更被正确记录到版本库。你不修改代码,只做版本控制操作。.
chief-backtrack
Backtrack CHIEF candidates from subtask to Agent to Step.