The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses local shell scripts to manage content and search files within the vault.
Evidence: scripts/init-content-item.sh and scripts/search-materials.sh utilize standard tools like mkdir, cp, rg (ripgrep), and grep to perform file operations and indexing.
[DYNAMIC_EXECUTION]: Template rendering is performed by executing the perl interpreter to modify file content.
Evidence: The render_template function in scripts/common.sh uses perl -pi -e to replace placeholders (e.g., {{TITLE}}) with values provided by the agent via environment variables.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface area for indirect prompt injection as it ingests untrusted data from web searches and research files to drive its workflow.
Ingestion points: The agent reads persona data, competitor analysis, and external search results as defined in SKILL.md (Step 3).
Boundary markers: The skill does not explicitly use delimiters or warnings to differentiate between system instructions and data ingested from external sources.
Capability inventory: The skill can create and write files (scripts/init-content-item.sh) and perform searches (scripts/search-materials.sh).
Sanitization: Filenames derived from user input are sanitized via the safe_title function in scripts/common.sh to prevent directory traversal or malformed paths.

jacky-content