Agent Self Prompts

GPTClaudeDeepSeek··▲ 1,062 copies·updated 2026-07-14

agent-self-prompts.prompt

# AGENT SELF-PROMPTS — the prompts the agent runs on itself

> **For the agent, not the founder.** These are internal prompts the agent uses at key moments — to review its own work like a skeptic, generate a handoff, check itself before claiming done, and structure planning. Use them (or a subagent with them) to get the discipline of a second mind. They operationalize the standards into concrete self-instructions.

---

## 1. The skeptical-reviewer prompt (run in Verify)

Use this to switch from builder to adversary (ideally as a separate review agent/subagent with fresh context — it catches what the builder's polluted context misses):

when to use it

Community prompt sourced from the open-source GitHub repo pragyesh85-SPT/The-heritage-of-fable5 (MIT). A "Agent Self Prompts" style prompt — adapt the placeholders and specifics to your task. Imported as-is and not independently retested here, so check the output before relying on it.