The Agent Skills Directory

[PROMPT_INJECTION]: The skill's instructions are focused entirely on fitness domain tasks. There are no attempts to bypass safety filters, extract system prompts, or override agent behavior.
[DATA_EXFILTRATION]: No network-enabled tools or commands (such as curl, wget, or fetch) are used. The skill handles personal fitness data locally within the context of the user's session and does not attempt to send information to external servers.
[REMOTE_CODE_EXECUTION]: No evidence of remote script execution, package installation, or dynamic code execution was found in the markdown instructions or the packaged skill file.
[SAFE]: The skill possesses a surface for indirect prompt injection by processing untrusted user workout data to generate social media updates and structured reports.
Ingestion points: User-provided workout notes, session logs, and PR updates (processed via the SKILL.md instructions).
Boundary markers: Absent; user inputs are processed as natural language without the use of explicit delimiters or guardrail prompts.
Capability inventory: Generates Facebook posts and structured exports (JSON-like, Markdown, and text reports).
Sanitization: None; the skill relies on the LLM to adhere to specific tone and formatting rules (e.g., 'no emojis', 'no em dashes') when interpreting user input.

fitness-log