The Agent Skills Directory

[SAFE]: No malicious code, obfuscation, or unauthorized access patterns were identified in the skill instructions.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from local project files which could potentially contain embedded instructions if those files are sourced from untrusted contributors.
Ingestion points: Reads from docs/sdd/{change-name}/proposal.md and .md files in ai-context/features/ (SKILL.md).
Boundary markers: Not explicitly defined in the provided manifest.
Capability inventory: Includes Read and Write operations for local documentation and mem_save for memory persistence.
Sanitization: No explicit sanitization of the content read from Markdown files is described in this wiring layer.

sdd-spec