The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data by reading existing artifact files from the disk to serve as context for generating subsequent planning documents.
Ingestion points: Files identified as dependencies in the openspec instructions JSON output are re-read from the local filesystem during the artifact creation loop (Step 5a).
Boundary markers: The skill establishes a "Planning boundary" instruction to limit the agent's scope to non-executable artifacts, though it does not explicitly define delimiters for the content of the files it reads.
Capability inventory: The skill possesses the capability to execute openspec CLI commands and write generated markdown or specification files to local paths.
Sanitization: No specific sanitization or filtering of the ingested file content is mentioned before it is processed by the agent.
[DYNAMIC_EXECUTION]: The skill documentation includes a mechanism where the agent may invoke specific commands or additional skills if the instruction field returned by the openspec instructions CLI tool delegates artifact creation to them. This represents a dynamic execution path where the agent's actions are guided by the output of the integrated CLI tool.

openspec-propose