The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and follow instructions from external sources at runtime, creating a significant attack surface for indirect prompt injection.
Ingestion points: In Step 1, the skill runs openspec instructions archive --change "<name>" --json and in Step 4, it runs openspec instructions specs --change "<name>" --json to fetch external data.
Boundary markers: The skill contains weak boundaries, instructing the agent to 'Treat context as a required prompt-level input' and 'apply relevant project facts, conventions, and constraints,' while only suggesting reporting conflicts rather than strictly isolating the data.
Capability inventory: The agent has access to Bash(openspec:*) tools, which include capabilities to write specs, archive changes, and perform file system operations like mkdir and mv.
Sanitization: There is no evidence of sanitization, validation, or filtering of the content returned by the CLI commands before it is adopted as behavior-modifying guidance.
[COMMAND_EXECUTION]: The skill performs shell-level operations, including directory creation (mkdir) and moving files (mv), as well as executing various subcommands of the openspec CLI. These operations are within the stated scope of the skill's purpose (archiving changes).

openspec-archive-change