Skill: empirical-proof

Purpose

Eliminate hallucinated completion. Coding agents are pattern-completers; the pattern of "successful task" includes confident-sounding completion language ("✅ all tests pass") that the agent will produce regardless of whether the underlying claim is true. Empirical proof is the structural defence: the agent cannot complete the pattern without first pasting evidence the pattern is true.

Project context (the AGENTS.md contract)

Resolves project commands via the consuming repo's AGENTS.md — the cmdValidate and cmdTest slots in its Commands table, plus an optional dep-validation command (architectural-rules check) not in the standard contract; ask the user if the project uses one. If AGENTS.md is missing or an entry is undefined, ask the user which command to run before pasting any "verification output" — do not guess.

Core rules

1. Never assume success

Writing the code is 10% of the job. Verifying it works in the current environment is 90%. The agent does not get to assume the test command's output without running it.

2. Verbatim pasting

When filling out ## Self-review, paste the verbatim output. No paraphrasing, no summarising, no "✅ passing". Use a fenced code block. Include the last two lines (or more if asked) — the runner's summary plus its timing/exit conditions.