The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes content from the wiki and raw data directories to identify contradictions, missing links, and metadata gaps.
Ingestion points: Reads content from wiki/ and raw/ directories, including markdown pages and source files, as specified in the checklist and README.md examples.
Boundary markers: Absent. The instructions do not include delimiters or warnings for the agent to ignore potentially malicious instructions embedded within the wiki content it audits.
Capability inventory: The skill has permissions to create and modify markdown files (e.g., creating stub pages, updating frontmatter, logging to wiki/log.md) and executes local CLI tools like qmd and grep.
Sanitization: Absent. There is no evidence of sanitization or validation of the ingested content before it is processed or used to generate summaries and log entries.

wiki-lint