@milpa/comal@0.2.1
Comal: governed operations, execution receipts, and programming assistance in any language
AI Security Review
scanned 4d ago · by lpm-firewall-aiThe flagged credential flows are provider requests constrained to the selected service origin, and the Unicode finding points to code that strips invisible controls. The reviewed entry points and relay operations show no concrete malicious behavior.
Decision evidence
public snapshot- The manifest exposes CLI binaries and declares no lifecycle scripts.
- The CLI sends subscription credentials only to the selected provider's origin; it rejects a different origin.
- Provider profiles require an explicitly configured HTTP(S) base URL and store environment variable names rather than API keys.
- The open-relay client restricts operations to listed reads and mutations requiring a human confirmation card.
- The flagged invisible Unicode characters are in a sanitizer that replaces control and bidirectional characters.
- The package uses network access and environment variables to connect to configured AI providers; this matches its CLI purpose.
- The open-relay plugin starts a watcher when activated unless its notify setting is off.
Source & flagged code
7 flaggedPackage source references a known benign dynamic code generation pattern.
node_modules/ajv/dist/compile/jtd/serialize.jsView on unpkg · L49Package source references dynamic require/import behavior.
node_modules/ajv/dist/vocabularies/dynamic/recursiveAnchor.jsView on unpkg · L2Findings
Affected versions and remediation
This report applies to @milpa/comal@0.2.1.
See all versions of @milpa/comal for other recorded verdicts.
- Review the evidence and your use of @milpa/comal@0.2.1 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.