The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill analyzes data from local files which likely contain external social media content.
Ingestion points: The skill reads from ~/.claude/x/voice.md, log.md, swipe.md, and ../x-post/hooks.json in SKILL.md to inform its planning.
Boundary markers: The instructions lack delimiters or specific warnings to the agent to treat ingested file content as untrusted data or to ignore embedded instructions.
Capability inventory: The skill can execute a local Python script (weights.py) and write plan output to ~/.claude/x/plan.md.
Sanitization: No filtering, escaping, or validation is performed on the content of the ingested markdown and log files before processing.
[EXTERNAL_DOWNLOADS]: Fetches public algorithm ranking parameters and commit history from the official X (Twitter) GitHub organization.
Evidence: The script weights.py retrieves source code from raw.githubusercontent.com/xai-org/x-algorithm/ and commit metadata from api.github.com/repos/xai-org/x-algorithm/ to provide the agent with current ranking weights.

x-plan