The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external markdown files located in the $WK_SKILLS_HOME/learnings/ directory and incorporates that content directly into the 'Common Mistakes' section of the newly created skill.
Ingestion points: Markdown files scanned via find and grep in Step 3 of SKILL.md.
Boundary markers: No explicit delimiters or instructions are used to separate the interpolated content from the skill's primary instructions.
Capability inventory: The skill possesses the ability to write to the file system, execute shell commands, and register new agent skills via npx.
Sanitization: Content from the external learning files is placed into the new skill without any validation, filtering, or escaping.- [COMMAND_EXECUTION]: The skill performs various file system and environment management tasks using shell commands.
Evidence: Uses mkdir -p to create directories, test for environment checks, and find/grep to search through local files.- [DYNAMIC_EXECUTION]: The skill generates new executable instruction files and associated metadata at runtime.
Evidence: Step 6 involves programmatically writing the complete content of a new SKILL.md and README.md to the local filesystem.- [EXTERNAL_DOWNLOADS]: The skill uses a package manager to verify and install the generated skill.
Evidence: Executes npx skills add in Step 8 to register the newly created skill with the agent environment.

wk-skill