Cloudflare Pages Function for Conduit Launch Jam entry intake and management. Handles jam submission validation, persistence to D1 database, anti-bot verification via Cloudflare Turnstile, and optional confirmation emails via Resend.
Single tool with well-structured schema and comprehensive parameter validation. Input schema is complete with proper types, descriptions, and constraints (maxLength, enum). Tool description is clear and action-oriented. However, output schema is not documented, LLMs cannot predict response structure. Error messages are user-friendly but lack recovery guidance. No tool annotations (destructiveHint, idempotentHint) despite WRITE risk classification.
POST /api/jam/submit — Conduit Launch Jam entry intake. Accepts jam submission with validation, anti-bot verification, and optional confirmation email.
Output schema not documented. Tool description states 'Accepts jam submission with validation...' but does not specify what the response contains. LLMs cannot plan downstream actions or extract confirmation details.
No tool annotations despite WRITE risk. Tool modifies state (UPSERT to database) but lacks destructiveHint or idempotentHint. LLMs cannot reason about retry safety or side effects.
Error messages lack recovery guidance. Errors like 'Could not save your entry. Please try again.' do not tell the LLM whether to retry, ask the user, or escalate. Per pattern:recovery-guide, errors should guide next steps.
Turnstile token exposed as required parameter. While the code does not log it, best practice is to inject anti-bot secrets server-side and never expose them as tool parameters. Agents log all parameters.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 51 | 2026-07-28+ | v2 |
No confirmation or dry-run pattern for irreversible operation. Submission is persisted immediately (UPSERT). Per pattern:confirmation-request, agents should confirm before executing state-modifying operations.