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 skills/tinyfish-io/tinyfish-cookbook/dep-securitynpx skills add tinyfish-io/tinyfish-cookbook --skill dep-securitygit clone --depth 1 https://github.com/tinyfish-io/tinyfish-cookbookWrote 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/skills/tinyfish-io/tinyfish-cookbook/dep-security)<a href="https://agentmods.dev/skills/tinyfish-io/tinyfish-cookbook/dep-security"><img src="https://agentmods.dev/badge/skills/tinyfish-io/tinyfish-cookbook/dep-security.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 | $0.00159 | $0.02463 |
| Opus 5 | $0.00079 | $0.01231 |
| Sonnet 5 | $0.00032 | $0.00493 |
| Haiku 4.5 | $0.00016 | $0.00246 |
Grade A, and why
dep-security 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 — 253 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Dependency Security Checker
Given a package.json, check every dependency against live CVE databases and security advisories — focusing on the last 48 hours, the window that cached tools miss.
Pre-flight check
tinyfish --version
tinyfish auth status
If not installed: npm install -g tinyfish
If not authenticated: tinyfish auth login
Step 1 — Parse dependencies
Read the package.json the user provided. Extract all package names and versions from:
dependenciesdevDependenciespeerDependencies(if present)
Produce a flat list: [{name, version, type}]
If the user hasn't provided a package.json, ask for it before proceeding. Do not guess.
Get today's date. Calculate the cutoff timestamp: now minus 48 hours. You will use this to filter results in every agent below.
Step 2 — Batch packages
Do not fire one agent per package — that would be extremely slow for large projects.
Instead batch into groups of up to 10 packages per agent and search for all of them at once. For a typical package.json with 20–40 deps, this means 2–4 agents total per source, all running in parallel.
Format each batch as a comma-separated search string:
express,lodash,axios,react,webpack etc.
Step 3 — Parallel security scan
Fire all agents simultaneously using & + wait. Each agent searches one source for all batches at once.
# ── BATCH SETUP ──────────────────────────────────────────────
# Split your package list into batches of 10, e.g.:
# BATCH_1="express,lodash,axios,react,next"
# BATCH_2="webpack,typescript,eslint,jest,prisma"
# (add more batches as needed)
TODAY=$(date +%Y-%m-%d)
# ── CVE DATABASE ─────────────────────────────────────────────
# One agent per batch, all in parallel
tinyfish agent run \
--url "https://cve.mitre.org/cgi-bin/cvekey.cgi?keyword={BATCH_1}" \
"You are on a CVE search results page. Today is {TODAY}.
You are looking for CVEs related to these npm packages: {BATCH_1}.
Scan ALL visible results on this page.
For each CVE that matches one of these package names AND was published or modified within the last 48 hours:
- CVE ID
- Package name it affects
- Severity (if shown)
- One-sentence description of what the vulnerability does
- Publication date
STRICT RULES:
- Do NOT click any CVE link
- Do NOT paginate
- Only include results from the last 48 hours — ignore older ones
- If nothing is within 48 hours, return an empty array
Return JSON array: [{cve_id, package, severity, description, published_date}]" \
--sync > /tmp/ds_cve_1.json &
# Repeat for each batch:
tinyfish agent run \
--url "https://cve.mitre.org/cgi-bin/cvekey.cgi?keyword={BATCH_2}" \
"You are on a CVE search results page. Today is {TODAY}.
You are looking for CVEs related to these npm packages: {BATCH_2}.
Scan ALL visible results on this page.
For each CVE that matches one of these package names AND was published or modified within the last 48 hours:
- CVE ID, package name, severity, one-sentence description, publication date
STRICT RULES:
- Do NOT click any CVE link — read the listing text only
- Do NOT paginate
- Only include results from the last 48 hours
Return JSON array: [{cve_id, package, severity, description, published_date}]" \
--sync > /tmp/ds_cve_2.json &
# ── GITHUB SECURITY ADVISORIES ───────────────────────────────
tinyfish agent run \
--url "https://github.com/advisories?query=ecosystem%3Anpm&order=newest" \
"You are on the GitHub Security Advisories page filtered to npm, sorted by newest first. Today is {TODAY}.
Read through the visible advisory listings on this page.
For each advisory that:
1. Affects any of these packages: {ALL_PACKAGES}
2. Was published within the last 48 hours
Extract:
- Advisory ID (GHSA-...)
- Package name
- Severity (Critical / High / Medium / Low)
- One-sentence description
- Published date
- Patched version (if shown in the listing)
STRICT RULES:
- Do NOT click any advisory to open it
- Do NOT paginate or click 'Load more'
- Scan only the first 30 visible listings then stop
- 48-hour cutoff is strict — ignore anything older
Return JSON array: [{ghsa_id, package, severity, description, published_date, patched_version}]" \
--sync > /tmp/ds_ghsa.json &
# ── NPM SECURITY FEED ────────────────────────────────────────
tinyfish agent run \
--url "https://www.npmjs.com/advisories" \
"You are on the npm security advisories page. Today is {TODAY}.
Read through the visible advisory listings.
For each advisory that:
1. Affects any of these packages: {ALL_PACKAGES}
2. Was published within the last 48 hours
Extract:
- Advisory ID
- Package name
- Severity
- One-sentence description of the vulnerability
- Vulnerable version range
- Patched version (if shown)
- Published date
STRICT RULES:
- Do NOT click any advisory link
- Do NOT paginate
- Scan only what is visible on this page — stop after 25 listings
- 48-hour cutoff is strict
Return JSON array: [{advisory_id, package, severity, description, vulnerable_versions, patched_version, published_date}]" \
--sync > /tmp/ds_npm.json &
# ── WAIT FOR ALL ─────────────────────────────────────────────
wait
echo "=== CVE BATCH 1 ===" && cat /tmp/ds_cve_1.json
echo "=== CVE BATCH 2 ===" && cat /tmp/ds_cve_2.json
echo "=== GHSA ===" && cat /tmp/ds_ghsa.json
echo "=== NPM ===" && cat /tmp/ds_npm.json
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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 · 253 lines · 159 tokens per session scan A 359889e31b0d
dep-security is a skill published in the GitHub repository tinyfish-io/tinyfish-cookbook (2,150 stars, last pushed 3d ago), licensed MIT. It adds 159 tokens to every session and 2,463 once invoked, about $0.0008 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 skills, from other repositories
smoke-test
End-to-end smoke test skill for DeerFlow. Guides through: 1) Pulling latest code, 2) Docker OR Local installation and deployment (user preference, default to Local if Docker network issues), 3) Service availability verification, 4) Health check, 5) Final test report. Use when the user says "run smoke test", "smoke…
image-generation
Use this skill when the user requests to generate, create, imagine, or visualize images including characters, scenes, products, or any visual content. Supports structured prompts and reference images for guided generation.
podcast-generation
Use this skill when the user requests to generate, create, or produce podcasts from text content. Converts written content into a two-host conversational podcast audio format with natural dialogue.
vercel-deploy
Deploy applications and websites to Vercel. Use this skill when the user requests deployment actions such as "Deploy my app", "Deploy this to production", "Create a preview deployment", "Deploy and give me the link", or "Push this live". No authentication required - returns preview URL and claimable deployment link.
skill-reviewer
Reviews DeerFlow skill packages for readiness, triggers, safety boundaries, resources, and evidence. Invoke when users ask to audit, grade, or production-check an existing skill.
surprise-me
Create a delightful, unexpected "wow" experience for the user by dynamically discovering and creatively combining other enabled skills. Triggers when the user says "surprise me" or any request expressing a desire for an unexpected creative showcase. Also triggers when the user is bored, wants inspiration, or asks for…