The Agent Skills Directory

[DYNAMIC_EXECUTION]: The script scripts/time_scan.py dynamically modifies the Python search path (sys.path) to import a module (dasha_pyjhora) from a relative directory path (../../vedic-calculator/scripts) that is calculated at runtime. This dynamic loading from computed paths is used to share functionality between different skills in the author's suite but presents a security risk by relying on the directory structure and runtime path resolution.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process significant amounts of untrusted data provided by the user, such as life events, personal traits, and family details. This data is used to drive the agent's reasoning and the generation of rectification reports.
Ingestion points: User-provided text descriptions of life events and personal traits; structured_data.md and user_context.md files.
Boundary markers: There are no explicit delimiters or specific instructions for the agent to treat user-provided event descriptions as untrusted data during analysis.
Capability inventory: The skill has the capability to write to the local file system (rectification_report.md, user_context.md) and execute a local Python script (scripts/time_scan.py).
Sanitization: The instructions do not specify any sanitization or validation of the natural language input provided by users regarding their life events.
[EXTERNAL_DOWNLOADS]: The requirements.txt file includes the ephem library, and the scripts/time_scan.py file references pyswisseph and pytz. While these are well-known and legitimate astronomical and time-handling libraries, the skill relies on their external availability and, in the case of pyswisseph, instructs the agent to probe the environment for an existing virtual environment containing the library.

vedic-rectifier