The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes content from untrusted external sources, including project specification files, design documents, and the source codebase, to generate its verification report.
Ingestion points: The skill reads data from contextFiles (specs, tasks, design) identified by openspec instructions apply and performs keyword searches across the codebase in SKILL.md steps 5, 6, and 7.
Capability inventory: The skill has the ability to execute commands via the openspec CLI tool, including archive, doctor, and sync specs as listed in the Store Selection section.
Sanitization: There is no evidence of sanitization or specific boundary markers used to prevent the agent from following instructions that might be embedded within the project files being analyzed.
Boundary markers: While the skill uses structured JSON output for tool results, it lacks explicit instructions to treat the content of the files as passive data.

openspec-verify-change