opencli-autofix

Pass

Audited by Gen Agent Trust Hub on Oct 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the Bash tool to execute opencli for diagnostics and gh for reporting issues to GitHub. These are core functionalities of the skill's workflow.
  • [DATA_EXFILTRATION]: The skill transmits diagnostic metadata (site, command, error code, and fix summary) to the jackwener/OpenCLI repository. This is a vendor-owned resource (matching the skill author) and is guarded by a mandatory instruction to ask for user confirmation before the transmission occurs.
  • [INDIRECT_PROMPT_INJECTION]: The skill parses trace artifacts (summary.md, trace-error.yaml) which may contain content derived from external websites.
  • Ingestion points: Reads trace-error.yaml, summary.md, and adapter source files.
  • Boundary markers: Includes 'Safety Boundaries' instructions and 'Scope constraints' that limit modifications to the adapterSourcePath identified in the trace.
  • Capability inventory: Includes Read, Edit, Write, and Bash capabilities used to modify and execute local adapter scripts.
  • Sanitization: Relies on structural analysis and specific error code checks (e.g., AUTH_REQUIRED) to prevent processing of sensitive states.
  • [DYNAMIC_EXECUTION]: The skill identifies, modifies, and subsequently executes local adapter code (.js or .ts files) via opencli. This is the primary intended purpose of the skill and is constrained to specific directory paths defined in the trace metadata.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 26, 2026, 09:03 AM