The Agent Skills Directory

[SAFE]: The skill implements a strictly instructional workflow for document preparation. It relies on local Markdown files (references/*.md) for domain-specific terminology and templates, ensuring consistent and professional output without external dependencies.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data in the form of tender documents (via $ARGUMENTS or user input).
Ingestion points: Tender documents provided as file paths or text blocks during project initialization.
Boundary markers: The instructions do not define explicit structural delimiters for the input data, but they require the agent to perform an initial analysis phase before generating content.
Capability inventory: The skill's capabilities are limited to directory creation and file writing for the generated bid documents. It lacks network access or shell execution tools.
Sanitization: The skill contains explicit 'Key Rules' that prohibit the fabrication of credentials, certifications, or data, and mandates a 20% human review process for all outputs to ensure accuracy and compliance.
[DATA_EXPOSURE]: No patterns for accessing sensitive system files (e.g., SSH keys, environment variables) or hardcoded credentials were detected. The skill's focus is entirely on project-specific tender documentation.
[REMOTE_CODE_EXECUTION]: There are no commands that download or execute remote scripts. References to OCR tools like 'MinerU' are provided as recommendations for the user to preprocess their own files.

bidwriter