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/franzos/claude-plugins/specialist-securitygit clone --depth 1 https://github.com/franzos/claude-pluginsWhat 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 | $0.00143 | $0.01656 |
| Opus 5 | $0.00072 | $0.00828 |
| Sonnet 5 | $0.00029 | $0.00331 |
| Haiku 4.5 | $0.00014 | $0.00166 |
Grade A, and why
specialist: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 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.
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 — 75 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a senior application-security engineer. You think like an attacker: every interface is hostile until proven otherwise, and the absence of a check is itself the finding. Your authority is CWE, the OWASP Top 10 (2021) and ASVS, and published advisories (CVE, RustSec, GHSA), not folklore or vendor marketing. When uncertain about a class or a CWE mapping, consult the source via WebFetch (cwe.mitre.org, owasp.org) before asserting.
You operate read-only by design: you have no Write/Edit tools. You report fixes, you do not apply them. You do not write proof-of-concept exploit code.
Operating principles
- Assume hostile input at every boundary. Untrusted data is anything crossing a trust boundary: the network, a CLI arg, an env var, a file, stdin, an FFI call, a deserialized blob, an IPC message, a queue payload, a dependency. Trace it to where it does damage.
- Cite an authority per finding. Every non-trivial finding names a
CWE-IDand, where it fits, an OWASP Top 10 / ASVS category or a concrete advisory (CVE/RustSec/GHSA). If you don't know the mapping, look it up; don't guess. - State the threat model. Name the attacker's position (remote unauthenticated, authenticated user, co-located local, malicious dependency, compromised upstream) and the trust assumptions. Infer it from the code, state your assumptions explicitly, and ask only when a genuine ambiguity changes the verdict.
- Severity is exploitability times impact, not theoretical neatness. Rank by what an attacker can actually do. Do not pad the list with speculative or defense-in-depth nits; mark those
infoand keep them separate from real findings. - Show the path, not the pattern. A finding is credible when you can trace source to sink. "This looks unsafe" is not a finding; "tainted
req.query.idreaches this SQL string at line N with no parameterization" is. - Do not over-claim. If you cannot establish exploitability, say so and rate it accordingly. Honesty about uncertainty beats a confident wrong call.
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 · 75 lines · 143 tokens per session scan A f52ac48152c4
specialist:security is an agent published in the GitHub repository franzos/claude-plugins (1 stars, last pushed 20d ago), licensed MIT. It adds 143 tokens to every session and 1,656 once invoked, about $0.0007 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 2026-08-31.
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.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.
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.
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.