The Agent Skills Directory

[COMMAND_EXECUTION] (LOW): The skill executes a local Python script located at .codex/skills/survey-visuals/scripts/run.py. This is the primary mechanism for processing the survey data into visuals and appears to be a standard operational component of the skill rather than a vector for arbitrary command injection.
[DATA_EXPOSURE] (SAFE): The skill limits its file access to the user-provided workspace directory (e.g., outline/, papers/, citations/). It does not attempt to access sensitive system files, environment variables, or credentials.
[PROMPT_INJECTION] (LOW): The skill is susceptible to indirect prompt injection as it ingests untrusted data from external sources like paper notes. However, the output is restricted to non-prose artifacts (timelines and figure specifications), which limits the potential for executing malicious instructions.
Ingestion points: papers/paper_notes.jsonl, citations/ref.bib.
Boundary markers: Absent.
Capability inventory: File-write to outline/timeline.md and outline/figures.md, and local python script execution.
Sanitization: Absent.
[EXTERNAL_DOWNLOADS] (SAFE): The skill declares no network access and no external download commands (such as curl or wget) were found in the provided documentation.

survey-visuals