The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by ingesting untrusted data from existing project files to generate or modify an execution plan.
Ingestion points: The skill reads PLAN.md and PRD.md to understand project context and determine augmentation needs (specified in SKILL.md Steps 1 and 2).
Boundary markers: The instructions do not mandate the use of delimiters or 'ignore-instructions' blocks when the agent reads these external files.
Capability inventory: The skill's primary capability is file-write (updating the execution plan markdown file). It does not have capabilities for network access, subprocess execution, or privileged system operations.
Sanitization: There is no requirement for the agent to sanitize or validate the content of the ingested markdown files before incorporating the information into the augmented plan.

augment-plan