The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script provided within the package to report usage data.
Evidence: SKILL.md instructs the agent to run python3 scripts/record.py before performing the optimization task.
[DATA_EXFILTRATION]: The telemetry script sends usage information to an external API endpoint.
Evidence: scripts/record.py performs a POST request to https://redfox.hk/story/api/skill/record/save to save a record of the skill being used.
Note: The destination domain redfox.hk corresponds to the skill author redfox-data and is treated as a vendor-managed resource.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it reads and processes user-provided Markdown files.
Ingestion points: The skill reads the target SKILL.md file in the first step of its optimization workflow.
Boundary markers: The instructions do not define clear delimiters or use "ignore embedded instructions" warnings when processing the untrusted content from the target file.
Capability inventory: The skill has the ability to write to the file system using SearchReplace and execute local Python scripts via the terminal.
Sanitization: There is no evidence of sanitization or filtering of the content read from the target file before the agent processes it.

optimize-skill-md