The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes markdown content from a local wiki, creating a surface for indirect prompt injection if the wiki contains instructions designed to override agent behavior.
Ingestion points: wiki markdown files accessed via scripts/ws.py grep and full file reads.
Boundary markers: The skill instructions mandate citing sources with wikilinks and explicitly reporting knowledge gaps, but do not provide specific delimiters for separating wiki content from agent instructions.
Capability inventory: Shell execution for search operations and local file writes for logging.
Sanitization: No explicit sanitization or escaping instructions are provided for handling search patterns or content interpolated into the logging command.
[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands to interact with the bundled scripts/ws.py helper and to perform logging operations.
Evidence: Instructions in SKILL.md provide shell command templates for python3 scripts/ws.py and printf >> log.md. The Python script itself is a read-only traversal tool with no external dependencies or additional subprocess capabilities.

ws-search