The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates the execution of local Python scripts (scripts/001_analyze_files.py and scripts/002_extract_content.py) to manage file discovery and content extraction. While these scripts perform standard file operations, they represent the execution of local code directed by the agent.
[INDIRECT_PROMPT_INJECTION]: The core functionality involves processing content from external Markdown files. If these files contain malicious instructions, they could potentially influence the agent's behavior during the analysis and report synthesis phases.
Ingestion points: Markdown files are read by 001_analyze_files.py and their content is stored in 002_content.yaml for agent processing in Phase 3.
Boundary markers: The workflow does not define explicit delimiters or instructions to the agent to treat the extracted content as untrusted data or to ignore embedded commands.
Capability inventory: The agent is tasked with "intelligent analysis" and "synthesizing" reports based on the extracted text, which involves interpreting and summarizing the input data.
Sanitization: Although the skill cleans Markdown artifacts like bold markers and images, it does not perform sanitization to detect or neutralize natural language instructions within the text.
[DATA_EXPOSURE]: The script 001_analyze_files.py recursively scans directories for all files ending in .md. If the search root is misconfigured or points to sensitive areas, the skill could expose information from private notes or configuration files formatted as Markdown.

obsidian-knowledge-filter