The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the mktemp system utility to generate a unique temporary file path. This is a standard and safe way to handle temporary file creation.
[DATA_EXPOSURE]: The skill processes conversation history and references existing project documentation (PRDs, ADRs, plans) to create a summary. This data access is necessary for the skill's primary purpose and does not involve unauthorized exfiltration.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the conversation history. While this presents a surface for indirect prompt injection (Category 8), the skill only summarizes the information into a text document and does not execute the content or use it to drive sensitive tool calls. The mandatory evidence chain is as follows:
Ingestion points: Current conversation history (referenced in SKILL.md instructions).
Boundary markers: None explicitly defined in the prompt.
Capability inventory: File writing via system tools and mktemp (SKILL.md).
Sanitization: None specified for the conversation summary.

handoff