Skills
Navigate
skills/opsmill/infrahub-skills/infrahub-reporting-skill-gaps/Gen Agent Trust Hub

infrahub-reporting-skill-gaps

Pass

Audited by Gen Agent Trust Hub on Oct 27, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the current conversation and past session logs to identify friction points. Ingestion points: reference.md specifies reading conversation and logs from ~/.claude/projects/. Boundary markers: absent. Capability inventory: Bash, Grep, Glob, WebFetch, infrahubctl, pytest. Sanitization: rules/evidence-no-customer-data.md mandates redaction of PII, hostnames, and secrets.
  • [COMMAND_EXECUTION]: The skill performs shell commands to gather environment info and search logs. Evidence: SKILL.md and reference.md describe running infrahubctl info, infrahubctl schema load, pytest, and grep.
  • [DATA_EXFILTRATION]: The skill accesses local session history in ~/.claude/projects/. While this involves sensitive data access, it is required for the skill's diagnostic purpose and is mitigated by strict redaction rules and a user-consent handoff workflow.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 27, 2026, 10:13 AM