Skills
skills/full-statck-skills/openspec-skills/openspec-config/Gen Agent Trust Hub

openspec-config

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill facilitates the configuration of the OpenSpec CLI tool. All analyzed commands, file paths, and configuration examples are legitimate and appropriate for the skill's intended purpose.
  • [INDIRECT_PROMPT_INJECTION]: The skill describes a configuration surface where external data from openspec/config.yaml is injected into the agent's context. Evidence: 1. Ingestion point: openspec/config.yaml. 2. Boundary markers: The skill instructions mention <project-context> and <project-rules> tags used to encapsulate injected content. 3. Capability inventory: Local command execution via the openspec CLI. 4. Sanitization: Not explicitly mentioned. This is a standard configuration mechanism for this tool and represents a low risk.
  • [COMMAND_EXECUTION]: The skill documents legitimate use of openspec config subcommands (list, set, get, unset, reset, edit). These commands are local to the user's environment and used specifically for tool maintenance.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 04:14 AM
Security Audit — agent-trust-hub — openspec-config