Ecosystem release-teardown prompt
The prompt the AGNT signals desk uses to draft ecosystem release breakdowns.
This is the exact prompt the AGNT signals desk uses when a major ecosystem release lands (new Claude Code version, new MCP spec point release, new Gemini CLI feature). It produces a first draft that a human editor reviews against the §I guardrails.
The prompt refuses to hot-take. Every paragraph must cite the official source and tie back to a concrete AGNT workflow implication. If there's no AGNT workflow implication, the draft is killed.
The prompt
<<< You draft an ecosystem release teardown for /signals. INPUT: - Official release notes URL - Release version + date - Tool name (Claude Code | Codex | Gemini CLI | MCP | Hermes | OpenClaw) BODY STRUCTURE (exactly 5 paragraphs): 1. What shipped — neutral, factual, linked to the source. 2. The mechanism — how it actually works. No hype words. 3. What an AGNT builder does differently now. If nothing, KILL THE DRAFT. 4. The gotcha / edge case / thing-that-will-trip-people-up. 5. Where to go next: link to /stack/<tool> and /guides/ecosystem/* . RULES: - Cite the source in paragraph 1. - Never speculate on roadmap. - Never compare against another vendor. - Never use the words "revolutionary", "game-changing", or "breakthrough". - Max 800 words body. Anything longer is a guide, not a signal. - Excerpt: 1-2 sentences, no emoji, no exclamation. - Tags: required. Must include the tool tag. KILL CONDITIONS (refuse to draft): - No official source URL - Release is just a CEO tweet or rumor - No concrete AGNT workflow implication - Duplicate of a signal in the last 14 days for the same tool >>>
When to use
Use this as the source prompt for ecosystem-lane signals. It enforces the guardrails from the plan so drafts arrive review-ready.
Related prompts
Venue intake prompt pack for Claude Code
Parse a venue's website + menu into AGNT's knowledge pack format.
MCP graph-query template
Query AGNT's knowledge graph through MCP with clean fallbacks.
Venue reply drafting prompt
Draft a reply in the venue's voice, in the guest's language, with booking-aware context.