The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes existing documentation within the project (e.g., in docs/adr/ or decisions/) to ensure new records match established conventions. While this involves reading local data that could theoretically contain instructions, the skill is constrained to writing documentation and lacks network tools, mitigating risk.\n
Ingestion points: Reads existing files from project directories like docs/adr/ (specified in SKILL.md Step 1).\n
Boundary markers: The instructions do not define specific delimiters for separating ingested content from instructions.\n
Capability inventory: The skill is limited to Read, Glob, Write, and Edit tools; no network access or command execution tools are available.\n
Sanitization: No content sanitization is performed on the ingested text.\n- [NO_CODE]: The skill consists exclusively of instructional Markdown and reference templates; it does not contain or execute any scripts, binaries, or external dependencies.

adr-author