The Agent Skills Directory

[SAFE]: The skill implements a link auditing pipeline for Hugo sites as described. It extracts links, calculates graph metrics (orphans, hubs, etc.), and validates link existence. All operations are consistent with the stated purpose.
[COMMAND_EXECUTION]: The skill uses tools like Grep and Bash for file system operations and link extraction. These are used in a controlled manner within the Hugo project directory.
[EXTERNAL_DOWNLOADS]: The scripts/link_scanner.py script performs network requests using the urllib library to validate external URLs found in the markdown content. This is an optional feature triggered by a flag and includes logic to handle known bot-blocking sites neutrally.
[DATA_EXFILTRATION]: No data exfiltration was detected. Network activity is limited to HTTP HEAD/GET requests for link validation, and no sensitive local data is sent to external servers.
[INDIRECT_PROMPT_INJECTION]: The skill processes markdown files provided by the user. While this represents a data ingestion surface, the logic is limited to extracting link patterns and generating a report. The instructions explicitly forbid the agent from modifying content files without user review, mitigating the risk of automated exploitation via injected instructions.

link-auditor