CyberEdge Monitor Public Code

Use cyberedge-agent; this machine bridge is not a human CLI.

Preconditions

• Require an existing Scope with a non-empty human authorization reference and at least one domain target.
• Require this exact Skill/version grant with scan.intelligence, task.read, finding.read, evidence.read, and report.read.
• Require the isolated provider adapter. The GitHub token belongs only to that adapter.

Workflow

1. Start policy_public_code_intelligence for the authorized Scope.
2. Watch the Task until it becomes terminal.
3. Correlate each detector=github-public-code Finding with its github.code.reference Observation and JSON Evidence.
4. Report the queried domain, repository, path, blob SHA, GitHub URL, Evidence ID, and lifecycle state.
5. Treat github.code.coverage as a completed negative evaluation for that exact domain.
6. Treat github.code.error or github.code.no_targets as unknown coverage. Do not resolve or suppress prior candidates from an error.

Every result is an Info review candidate. It proves only that GitHub code search returned public repository metadata referencing the domain. It does not prove that a secret, credential, exploit, or sensitive source is present.

Never fetch repository content, clone a returned repository, print or request the provider token, widen the query, search domains outside the Scope, or describe a candidate as a confirmed leak.

{"request_id":"req-1","idempotency_key":"idem-1","agent_id":"codex-main","skill_name":"cyberedge-monitor-public-code","skill_version":"0.1.0","action":"start_scan","scope_id":"scope-id","policy_id":"policy_public_code_intelligence"}