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 skills add verivus-oss/llm-cli-gateway --skill red-team-assessmentgit clone --depth 1 https://github.com/verivus-oss/llm-cli-gatewayWrote 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/verivus-oss/llm-cli-gateway/red-team-assessment)<a href="https://agentmods.dev/skills/verivus-oss/llm-cli-gateway/red-team-assessment"><img src="https://agentmods.dev/badge/skills/verivus-oss/llm-cli-gateway/red-team-assessment/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/verivus-oss/llm-cli-gateway/red-team-assessment"><img src="https://agentmods.dev/badge/skills/verivus-oss/llm-cli-gateway/red-team-assessment.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 5 findings, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Anti-Refusal · line 297 Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.Fix: Remove instructions that suppress warnings, disclaimers, or ethical commentary. Let the agent surface safety-relevant caveats to the user.
- medium Prompt Injection · line 81 Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.Fix: Remove the large whitespace padding (blank-line blocks or long space runs) and review any content hidden below or to the right of it. Keep skill files compact and reviewable so no instructions can be
- medium Prompt Injection · line 83 Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.Fix: Remove the large whitespace padding (blank-line blocks or long space runs) and review any content hidden below or to the right of it. Keep skill files compact and reviewable so no instructions can be
- medium Prompt Injection · line 84 Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.Fix: Remove the large whitespace padding (blank-line blocks or long space runs) and review any content hidden below or to the right of it. Keep skill files compact and reviewable so no instructions can be
- medium Prompt Injection · line 85 Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.Fix: Remove the large whitespace padding (blank-line blocks or long space runs) and review any content hidden below or to the right of it. Keep skill files compact and reviewable so no instructions can be
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 | $0.00065 | $0.03157 |
| Opus 5 | $0.00032 | $0.01579 |
| Sonnet 5 | $0.00013 | $0.00631 |
| Haiku 4.5 | $0.00006 | $0.00316 |
Grade A, and why
red-team-assessment 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 11d 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 — 300 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Red-Team Assessment
Use the local gtwy stdio MCP server for every red-team and blue-team request. Do not launch a provider CLI directly. Start by calling provider_tool_capabilities to verify the live provider, transport, tool, and target-access surface.
The complete gateway CLI roster is Claude, Codex, Gemini, Grok, Mistral, Devin, and Cursor. A full red-team assessment dispatches every required reviewer in that roster. If a required provider is unavailable, report a blocker and repair it or obtain explicit user direction. Do not silently call a smaller set "complete."
Configured API providers are discovered dynamically with list_models and
their reported capabilities. They do not provide a local CLI checkout/worktree
or native ACP boundary, so they are not silently interchangeable with a
required source-inspecting red-team reviewer.
Security Gate Contract
Every red-team prompt must finish with the exact terminal verdict schema:
APPROVED_UNCONDITIONALLY | CHANGES_REQUIRED | BLOCKED_EXTERNAL
APPROVED_UNCONDITIONALLY is allowed only when no unresolved finding or
verification gap remains. CHANGES_REQUIRED must include evidence-backed
findings. BLOCKED_EXTERNAL is reserved for a concrete external access,
provider, or environment failure with its exact error. Approval with caveats,
an accepted-but-unverified claim, a skipped surface, a malformed response,
timeout, cancellation, or provider failure is not unconditional approval.
Apply these rules:
- Omit model unless the caller explicitly selected it.
- Use approvalStrategy: "legacy" for Codex, Gemini, Grok, Mistral, Devin, and Cursor. mcp_managed is rejected for them and approvalPolicy has no effect.
- Use mcp_managed only for a deliberately configured Claude request. In that mode, request-scoped strict MCP configuration is limited to provisioned gateway-owned definitions; do not assume ambient research tools are present.
- Use sandboxMode: "read-only" for inspection-only Codex work, or sandboxMode: "workspace-write" only when red-team verification needs to create build or test artifacts. Do not use fullAuto.
- Do not set review-round, turn, token, price, cost, or wallclock caps. The
configured idle-timeout safeguard only detects a silent process on incremental providers. For terminal-burst providers (gemini, mistral, devin, cursor) it is instead a total-runtime bound, since those emit nothing until exit. That classification is each provider's default invocation; cursor with
outputFormat: "stream-json"streams, so there it is a real idle window. - Continue red-team, blue-team, and reassessment work until every required
reviewer returns
APPROVED_UNCONDITIONALLY.
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.
- 11d ago First seen · 300 lines · 65 tokens per session scan A 71f6635f4c18
red-team-assessment is a skill published in the GitHub repository verivus-oss/llm-cli-gateway (16 stars, last pushed yesterday), licensed MIT. It adds 65 tokens to every session and 3,157 once invoked, about $0.0003 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
architect
System design, tradeoffs, and complex technical decisions.
scope-analyst
Catch ambiguities and hidden requirements before planning.
security-analyst
Threat-model and find vulnerabilities, with practical remediation.
code-reviewer
Find bugs, security holes, and maintainability issues in a diff or file.
plan-reviewer
Validate that a work plan is executable before work starts.
researcher
Research external libraries, APIs, and best practices, with evidence.