council — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited council (Agent Skill) and scored it 91/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 1 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 1 flagged
The text {match} is the classic direct prompt-injection phrasing. Placed in a skill body that the agent reads as trusted instructions, it tries to make the agent abandon its prior rules and follow whatever comes next — a full system-prompt override.
ignore/disregard/forget … previous instructions sentence.Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Consult the cli-bridge council on the user's question.
cwdargument — never paste file contents into the task (the delegates read the files themselves; inlining wastes tokens).
ask_all MCP tool (server cli-bridge) with the question as task.Add synthesize: true when the user wants one merged answer rather than the side-by-side responses.
include_paid: true)only when the user asks for "the best" — their cost profile may refuse it.
synthesis: where the models agree, where they disagree, and what you'd do.
If ask_all reports no lanes installed, run the doctor tool and relay its install hints instead of failing silently.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.