Forge-core VerifyCompletion

Verify work before claiming done. USE WHEN about to claim work is complete, fixed, or passing, before committing or creating PRs.

install
source · Clone the upstream repo
git clone https://github.com/N4M3Z/forge-core
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/N4M3Z/forge-core "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/VerifyCompletion" ~/.claude/skills/n4m3z-forge-core-verifycompletion && rm -rf "$T"
manifest: skills/VerifyCompletion/SKILL.md
source content

VerifyCompletion

No completion claims without fresh verification evidence. Run the command, read the output, then claim the result.

The Gate Function

Before claiming any status or expressing satisfaction:

  1. Identify the command that proves the claim
  2. Run the full command now (not from memory, not from a previous run)
  3. Read the complete output, check exit code, count failures
  4. Match the output to your claim — does it confirm what you're about to say?
  5. Only then make the claim, citing the evidence

Skip any step and the claim is unverified.

Verification Requirements

ClaimRequiresNot sufficient
Tests passTest command output showing 0 failuresPrevious run, "should pass"
Linter cleanLinter output showing 0 errorsPartial check, extrapolation
Build succeedsBuild command exit 0Linter passing, "logs look good"
Bug fixedReproducing original symptom: now passesCode changed, assumed fixed
Requirements metLine-by-line checklist against specTests passing
Agent completedVCS diff shows expected changesAgent reports "success"

Red Flags

ThoughtReality
"Should work now"Run the verification.
"I'm confident"Confidence is not evidence.
"Just this once"No exceptions.
"Linter passed"Linter is not the compiler. Run the build.
"Agent said success"Verify independently. Agents report intent, not fact.
"Partial check is enough"Partial proves nothing about the whole.
"I already tested this earlier"Earlier is not now. State changes between runs.
"It's a small change, can't break"Small changes break large systems. Verify.

Constraints

  • Never express satisfaction ("Done!", "Perfect!", "All good!") before running verification
  • Never trust a previous run — re-run now
  • Never trust agent success reports without checking the diff
  • "Should", "probably", "seems to" are never verification