The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized network behaviors were detected. The skill follows best practices for data integrity by requiring manual verification of automated findings.
[COMMAND_EXECUTION]: The skill invokes a local script scripts/observations.py to process image files and generate scaled crops. This operation is limited to local relative paths and is a documented part of the skill's primary function.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data sources which represents a theoretical injection surface. 1. Ingestion points: The skill reads project records from source-observations.md and processes external photographic and scan data. 2. Boundary markers: No explicit prompt delimiters are defined. 3. Capability inventory: The skill uses scripts/observations.py to read source files and write annotated crops to the file system. 4. Sanitization: The skill implements manual review as a security control, requiring the user to 'confirm the full mark lies on the claimed construction' and forbidding the synthesis of 'accepted' statuses from author labels or automated loops.

source-observe