strong agent prompts for bamf api

strong agent prompts are your key to leveraging the BAMF API effectively. start with read-only proof, pick your creator space, and preview scheduled content before making any changes. approvals matter—no mutations without consent. here’s a nuts-and-bolts guide to help you in building a small BAMF API client. first off, make sure your BAMF API key is stored as an environment variable, BAMF_API_KEY. avoid hardcoding it; keep it secure. ``` You are building a small BAMF API client for me. Use my BAMF API key only as a local environment variable named BAMF_API_KEY. Do not hardcode it, print it, or commit it. REST base URL: https://api.bamf.ai/v1 MCP URL: https://mcp.bamf.ai First: 1. Call GET /agent/capabilities to load the current workflow map. 2. For ambiguous requests, call POST /agent/route with my natural-language request. 3. Call GET /me to verify auth. 4. Call GET /creator-spaces and ask me which creator_id to use if more than one is available. 5. Call GET /posts?status=scheduled&include=media,first_comment,preview. 6. Read ideas, knowledge, media, and analytics before drafting when scopes allow it. 7. Create one draft post only after I confirm the exact content. Use an Idempotency-Key header for every POST, PATCH, or DELETE. Never publish-now unless I explicitly ask for it. ``` using this structured approach, you’ll get to the right content faster, all while reducing mistakes. keep that Idempotency-Key in mind—your safety net for creating, updating, or deleting content. stay focused. your creator workflow deserves precision and clarity. originally from [BAMF.ai](https://bamf.ai/docs/examples/agent-prompts).