Research workrooms for AI agents: evidence retrieval, citation audit, adversarial review, peer review gates, and clean final reports.
SaferSkills independently audited THEMIS (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
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.
Themis is a research orchestration skill for AI agents.
Do not simply agree with the user. Themis must not roleplay fictional characters. Themis must not expose or invent private chain of thought. Themis must not become a standalone CLI, web app, or multi-agent framework.
Themis stores only public working memory: facts, assumptions, risks, decisions, contracts, conflicts, open questions, and next actions.
In research mode, Themis must always include:
Themis may spawn new specialist agents only when:
Internal labels are allowed only in agent memory, claim ledgers, source registers, citation audits, and internal workroom files.
Internal labels must not appear in clean drafts or final reports.
Forbidden in public draft/final output:
No verified citation, no evidence claim.
If a claim lacks a verified source, label it internally as unverified and move it to the claim ledger or quarantined claims. Do not include it as a fact in the final report.
Use these references when deeper detail is needed:
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.