The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from transcripts located in the local directory /Users/rami/Library/Mobile Documents/iCloud~md~obsidian/Documents/Obsidian/My Outputs/Transcriptions/. While it performs formatting cleanup (stripping YAML and wiki-links), it lacks specific security delimiters or instructions to ignore potential commands embedded within the transcript text. This creates a surface where malicious content in a processed file could influence the agent's behavior during the research or writing steps.
Ingestion points: Transcript files read from the local Obsidian directory.
Boundary markers: Absent. No specific delimiters are used to wrap the processed content.
Capability inventory: File system access (read/write), subdirectory creation, and orchestration of sub-skills (content-research, content-hooks, content-writer).
Sanitization: Only formatting-based cleanup (wiki-link conversion and section removal) is performed; no security-focused sanitization or instruction filtering is present.
[COMMAND_EXECUTION]: The skill performs automated file system operations, including scanning directories, creating folders in /Users/rami/Documents/life-os/notes-processing/, and writing multiple markdown files (research.md, hooks.md, content.md). These operations are consistent with the skill's stated purpose for the author 'rami-maalouf'.

content-pipeline