The Agent Skills Directory

[COMMAND_EXECUTION]: Uses shell commands to interact with the git repository and modify files.
Employs git rev-parse and sources a local script .specify/scripts/bash/common.sh for environment context.
Modifies files by appending content via cat and heredocs.
[PROMPT_INJECTION]: Vulnerable to indirect prompt injection through external document processing.
Ingestion points: Content is read from a user-supplied <document-path> (file SKILL.md).
Boundary markers: Utilizes quoted heredocs ('EOF') to safely append content to files, but lacks explicit constraints to prevent the agent from following instructions embedded in the source documents.
Capability inventory: Able to modify project files and invoke other high-level workflow commands like /speckit.plan and /speckit.tasks (file SKILL.md).
Sanitization: Content from external documents is processed and incorporated without validation or filtering for malicious instructions.

speckit-incorporate