The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from multiple sources, creating a surface for indirect prompt injection where instructions hidden in external documentation or codebase files could influence the agent's output.
Ingestion points: The skill uses WebSearch and WebFetch to research framework documentation and patterns. It also uses Read, Glob, and Grep to ingest arbitrary files from the local codebase (e.g., openspec/project.md, proposal.md, design.md).
Boundary markers: The instructions do not define clear boundaries or provide specific guidance for the agent to ignore instructions embedded within the ingested data.
Capability inventory: The skill possesses Edit and Write capabilities, allowing it to generate or modify project artifacts (tasks.md, tests.md, proposal.md) based on the data it reads.
Sanitization: There is no explicit requirement for the skill to sanitize or validate the content retrieved from the web or project files before interpolation into its planning logic.

openspec-plan