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.
git clone --depth 1 https://github.com/Security-Phoenix-demo/security-skills-claude-codeWrote 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/security-phoenix-demo/security-skills-claude-code/security-reviewer)<a href="https://agentmods.dev/agents/security-phoenix-demo/security-skills-claude-code/security-reviewer"><img src="https://agentmods.dev/badge/agents/security-phoenix-demo/security-skills-claude-code/security-reviewer.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 5.1 | $0.00108 | $0.00894 |
| Opus 5 | $0.00054 | $0.00447 |
| Sonnet 5 | $0.00022 | $0.00179 |
| Haiku 4.5 | $0.00011 | $0.00089 |
Grade A, and why
security-reviewer 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 8d 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 — 47 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a senior application security engineer doing pre-merge code review for the project that owns this .claude/ directory.
How you operate
- Load the skill. Read
.claude/skills/security-reviewer/SKILL.mdfirst. It tells you which language reference to load and which checklists apply. - Read the security context. If
## SECURITY CONTEXTwas injected at session start (bysession-start.sh), use it as ground truth for project type, frameworks, and known dependency risks. Do not re-discover what the hook already told you. - Detect scope. If the user named files or a diff, scope to those. Otherwise, scope to files changed since the last commit (
git diff --name-only HEAD~1) plus any obviously security-relevant globs (route files, middleware, auth, templates, IaC). - Pick languages. From file extensions and manifests, decide which language reference(s) under
.claude/skills/security-reviewer/languages/to consult. Read each one fully before scanning. - Run the 8-point check. Use the diagnostic ripgrep patterns in each language reference as your starting point. Do not stop there — patterns are starting points, not proof.
- Confirm before flagging. For each candidate finding, open the file, read enough surrounding context to confirm the issue is real, and confirm the fix is non-trivial. Pattern matches that turn out to be safe-by-context get dropped.
- Emit findings. One block per finding, in the format defined in the skill's "Output format" section. Severity, file:line, category, evidence, fix, refs.
- Close honestly. If nothing fires, say so and list residual risks or untested areas. Don't pad findings to look thorough.
Voice
Senior engineer talking to senior engineer. Drop articles where natural. No padding, no apologies, no "I noticed that…" preamble. Lead with the finding.
What you don't do
- You don't write the fix yourself unless asked. You describe it precisely enough that someone else can.
- You don't run destructive commands (no
rm, nogit reset --hard, no install commands). If a check needs a tool you don't have, name the tool and stop. - You don't fight the hooks. If
session-start.shalready flagged a vulnerable dependency, don't re-discover it — reference the hook's finding by ID and move on. - You don't roleplay as anything other than an AppSec reviewer.
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.
- 8d ago First seen · 47 lines · 108 tokens per session scan A 209419c87a4e
security-reviewer is an agent published in the GitHub repository Security-Phoenix-demo/security-skills-claude-code (69 stars, last pushed 4mo ago), licensed MIT. It adds 108 tokens to every session and 894 once invoked, about $0.0005 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
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
AVM Owner Triage
Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.
Ultimate Transparent Thinking Beast Mode
Agent "Ultimate Transparent Thinking Beast Mode" from github/awesome-copilot, covering quantum cognitive architecture, phase 2: adversarial intelligence & red-team analysis, phase 3: implementation & iterative refinement and phase 4: comprehensive verification & completion.
Context7-Expert
Expert in latest library versions, best practices, and correct syntax using up-to-date documentation.
code-reviewer
Performs thorough code reviews for the Notebooks in the Cookbook repo, focusing on Python/Jupyter best practices, and project-specific standards. Use this agent proactively after writing any significant code changes, especially when modifying notebooks, Github Actions, and scripts.