The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses curl commands to upload user files and submit task payloads to an internal service located at http://opencode-bridge:8000. This is used for delegating document generation tasks.
[EXTERNAL_DOWNLOADS]: The skill communicates with a local network bridge (opencode-bridge). While this service appears to be part of the skill's infrastructure, it involves transmitting user-provided data and authentication tokens over the network.
[PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes untrusted external data.
Ingestion points: Phase 2 allows the upload of arbitrary user files including Word docs, PDFs, spreadsheets, and logs into the input/ directory.
Boundary markers: The task submission template in Phase 3 does not use explicit delimiters or instructions to ignore embedded commands within the referenced input files.
Capability inventory: The skill uses curl to interact with an execution bridge (opencode-bridge) which performs complex document generation.
Sanitization: No explicit sanitization, validation, or content filtering of the uploaded user files is performed before they are processed by the backend service.

tech-design-writer