The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local shell scripts (validate-plan.sh and validate-all-plans.sh) to perform content validation and ensure plan quality.
[COMMAND_EXECUTION]: The script validate-all-plans.sh modifies file system permissions for the validator script using the chmod +x command.
[COMMAND_EXECUTION]: The validation logic relies on common shell utilities including find, grep, and sed for directory traversal and text analysis.
[PROMPT_INJECTION]: The skill facilitates indirect prompt injection by ingesting and processing content from the project's codebase to generate strategic plans.
Ingestion points: Codebase content is retrieved through search, sem_search, and read tools as described in SKILL.md.
Boundary markers: While the generated plans use specific Markdown headers, there are no explicit delimiters or instructions provided to the agent to treat codebase content as untrusted or to ignore embedded instructions.
Capability inventory: The skill environment permits shell command execution (chmod, find, grep, sed) and file system searches.
Sanitization: The skill does not implement sanitization or filtering for the content ingested from the codebase before it is used in prompt generation.

create-plan