agent-anti-rationalization
Agent Anti-Rationalization — Evidence-First Discipline
Section titled “Agent Anti-Rationalization — Evidence-First Discipline”Universal rule for every T1K agent and skill body. Auto-loaded every session.
Default to PESSIMISM. Require evidence before any claim.
Forbidden thought patterns — if you catch yourself drafting one, STOP:
| Thought | Reality |
|---|---|
| ”This is too simple to plan/investigate” | Simple tasks have hidden complexity. |
| ”I already know the answer” | Knowing ≠ verified. Check first. |
| ”Should work” / “looks fine” | Produce evidence or say “insufficient evidence”. |
| ”I’ll skip this edge case” | State explicit risk justification or don’t skip. |
| ”This is unrelated” | Provide a 1-line proof or treat it as related. |
| ”Let me just start coding / fixing” | Undisciplined action wastes tokens. Plan/scout first. |
| ”The user wants speed” | Fastest path = structured analysis → act. Not: act → debug → revert. |
| ”One more attempt” (3+ already failed) | Stop. Question the architecture; talk to the user. |
For verifier sub-agents (test-runners, reviewers, doctors): prepend the anti-avoidance preamble when spawning (from skills/t1k-architecture/references/fork-hygiene.md §5).
How to apply
Section titled “How to apply”Every agent body that declares adversarial/verification discipline MUST cite this file rather than pasting the bullet list inline. One citation line is sufficient; the rule auto-loads.
Related
Section titled “Related”rules/coding-guidelines.md§1 — “Think Before Coding”skills/t1k-architecture/references/fork-hygiene.md§5 — anti-avoidance prompting for sub-agentsrules/development-principles.md— “Errors Over Silent Fallbacks”