openspec-archive-change

Pass

Audited by Gen Agent Trust Hub on Aug 20, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill uses 'openspec instructions archive' and 'openspec instructions specs' to fetch dynamic context and guidance that the agent is then directed to follow as 'behavior contracts'. This creates a surface for indirect prompt injection where instructions embedded in external data could influence or override intended behavior.
  • Ingestion points: 'openspec instructions archive --change "" --json' and 'openspec instructions specs --change "" --json' (SKILL.md).
  • Boundary markers: The instructions specify that these fields should be kept separate from built-in steps and that conflicts must be reported.
  • Capability inventory: The skill uses the 'Bash' tool for command execution and file system operations such as 'mkdir' and 'mv'.
  • Sanitization: The agent is explicitly instructed not to copy the guidance text verbatim into outputs and to explain any inapplicable advice.
  • [COMMAND_EXECUTION]: The skill constructs shell commands by interpolating variables like 'change_name' and 'store_id' into bash strings. If these variables, which can be derived from project metadata or CLI outputs, contain malicious shell metacharacters, they could facilitate command injection attacks.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 20, 2026, 10:25 AM
Security Audit — agent-trust-hub — openspec-archive-change