The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted user data, specifically a career inventory, to generate resume drafts.
Ingestion points: Data enters the process via JSON input passed to scripts/check_resume_plan.py as described in SKILL.md and references/drafting-method.md.
Boundary markers: The skill uses structured JSON objects with mandatory field validation defined in scripts/_common.py (e.g., require_text, require_object).
Capability inventory: The skill's capabilities are limited to text processing, timeline calculation, and flagging inconsistencies. It lacks network access, arbitrary shell execution, or file system write capabilities.
Sanitization: scripts/_common.py includes regex patterns (EMAIL_PATTERN, PHONE_PATTERN, MYNUMBER_PATTERN) to identify and protect sensitive personal information within the processed text.
[COMMAND_EXECUTION]: The skill invokes a local Python script for logic validation.
Evidence: SKILL.md step 5 explicitly uses python3 scripts/check_resume_plan.py <input.json> to verify the resume plan.
Context: The script performs deterministic calculations on structured data (finding timeline gaps/overlaps) and enforces the vendor's data-handling policies (checking for unconfirmed metrics or un-abstracted confidential info).

resume-draft