The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands using the gh (GitHub CLI) to fetch repository metadata, pull requests, and commit history. It also runs a local Python script, skillctx-resolve.py, to manage and resolve configuration variables stored in the user's config directory.
[DATA_EXFILTRATION]: Accesses multiple sensitive data sources including local Claude Code session history files, GitHub pull request bodies and commit messages, Slack message threads, and meeting transcripts from Fireflies. This access is necessary for the skill's primary function of mining activity for blog topics but exposes significant personal and professional information to the agent.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external platforms (Slack, GitHub, Fireflies) which could contain malicious instructions designed to influence the agent's behavior during analysis or outline generation.
Ingestion points: Mined content from slack_search_public_and_private, GitHub pull request bodies, and Fireflies transcripts (via fireflies_get_transcripts).
Boundary markers: Absent; the skill does not define specific delimiters or instructions to ignore embedded commands in the mined data.
Capability inventory: Execution of gh CLI commands and writing generated outlines to the local filesystem at {blog_drafts_dir}.
Sanitization: While the instructions mandate redacting secrets and personal names for privacy, they do not include measures to sanitize or escape instructions embedded within the ingested data.

p-blog-post-mining