The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection because it reads and merges files from openspec/changes/ and openspec/specs/. Instructions embedded in these files by an attacker could potentially influence the agent's behavior during the merging or conflict resolution process.
Ingestion points: Local filesystem paths openspec/changes/ and openspec/specs/.
Boundary markers: There are no explicit instructions or delimiters provided to ensure the agent ignores embedded commands within the processed files.
Capability inventory: The skill involves filesystem write operations, including moving files to openspec/changes/archive/ and merging content into the specs/ directory.
Sanitization: The workflow requires the user to review and confirm the resolution plan before final execution, which provides a layer of human oversight.

openspec-bulk-archive