openspec-archive-change

Pass

Audited by Gen Agent Trust Hub on Mar 7, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands such as openspec status --change "<name>" and mv openspec/changes/<name> .... The <name> variable is sourced from user input or conversation context. Although the instructions use double quotes for interpolation, there is no explicit requirement to sanitize the input for shell metacharacters, which could lead to unintended command behavior if the change name is maliciously crafted.
  • [PROMPT_INJECTION]: The skill processes untrusted data from external sources, specifically the contents of tasks.md and the JSON output of the openspec CLI. This constitutes an indirect prompt injection surface.
  • Ingestion points: The agent reads the tasks.md file in Step 3 and parses CLI output in Step 2.
  • Boundary markers: There are no explicit delimiters or instructions provided to the agent to ignore or isolate potential natural language instructions embedded within tasks.md or the CLI output.
  • Capability inventory: The agent has the capability to create directories (mkdir), move files (mv), and execute project-specific CLI tools (openspec).
  • Sanitization: No sanitization or validation logic is specified for the content read from tasks.md before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 7, 2026, 09:04 AM