
Metal Mantra
Public evidence reports for AI agents
▤ Screenshots
About Metal Mantra
AI agents now act on your behalf, yet most agent directories are lists of pitches. Metal Mantra publishes evidence instead.
Paste a public GitHub repository and an official domain, and deterministic rules read the source and produce a public report:
- Five weighted pillars: security & privacy, guardrails & loop control, schema & tooling, token & cost efficiency, entity trust.
- Findings you can inspect, with the file and rule behind every deduction.
- Scan scope stated plainly: how many files were eligible, how many were read, and what that leaves unknown.
- Domain evidence, clearly marked when unverified.
Buyers can search the registry or describe a job and see agents whose public descriptions may fit. Builders get a badge for their README and can re-scan after fixes. Private repo? Run the scanner in your own GitHub Actions and publish an attested report without sharing code.
There is no AI in the scanner, so the same input gives the same result. A score is an automated signal about public source, never a security certification, and every report says what it does not establish. The first scan of a repo is free.
✦ Good to know
- The first supported public repository scan is free, and a re-audit pass is available after the free scan.
- Paid re-audit passes provide three re-scans for one repository, with no subscription or auto-renewal.
- Reports cover five pillars: security and privacy, guardrails, schema and tooling, token efficiency, and entity trust.
- Each report records findings, scan scope, timestamp and unknowns; it is not a certification or guarantee of runtime behavior.
Taken from the product's own website and re-checked automatically.
Still not sure about Metal Mantra?
Let your favourite AI do the thinking. Pick one and see what it says.
Reviews
Promoted
Upvote
More in AI
Upvote
Upvote
Upvote
Upvote
Upvote
Upvote
You may also like
Upvote
Upvote
Comments