Stats
Actions
Tags
From superpowers-plus
Dispatches acceptance-tester subagent to verify branch satisfies all acceptance criteria after code review and before finishing branch.
How this skill is triggered — by the user, by Claude, or both
Slash command
/superpowers-plus:acceptance-testingThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
**Announce at start:** "I'm using the acceptance-testing skill to verify the branch against acceptance criteria."
Announce at start: "I'm using the acceptance-testing skill to verify the branch against acceptance criteria."
Dispatch superpowers:acceptance-tester subagent to verify the completed branch satisfies every acceptance criterion before integration. The tester receives the AC document path and the current git HEAD SHA.
Core principle: Implementation is not done until every acceptance criterion has recorded evidence of a pass. A failing criterion is a blocker, not a review comment.
NO BRANCH FINISHING UNTIL ACCEPTANCE TESTING REPORTS ALL CRITERIA PASSED
(OR EXPLICITLY DEFERRED WITH USER APPROVAL).
Mandatory:
requesting-code-review completes and all Critical and Important issues are resolvedfinishing-a-development-branchIf no AC document exists:
Stop. Invoke superpowers:writing-acceptance-criteria to produce the document from the spec, get user approval, then return here.
1. Identify inputs:
AC_DOC="docs/superpowers/acceptance/YYYY-MM-DD-<feature>.md"
HEAD_SHA=$(git rev-parse HEAD)
REPO_ROOT=$(git rev-parse --show-toplevel)
2. Dispatch acceptance-tester subagent:
Use Task tool with superpowers:acceptance-tester type, fill template at acceptance-tester.md
Placeholders:
{AC_DOC_PATH} — path to the acceptance criteria document{HEAD_SHA} — current git HEAD SHA{REPO_ROOT} — absolute path to the repository root3. Act on the report:
| Report outcome | Action |
|---|---|
| All criteria PASS | Proceed to superpowers:finishing-a-development-branch |
| One or more criteria FAIL | Return to superpowers:executing-plans: create fix tasks for each FAIL criterion |
| Criteria Blocked due to failed dependency | Will unblock automatically once their dependency criteria pass; treat as FAIL for fix prioritization |
| Criteria Blocked due to missing infrastructure | Stop and report to user; get explicit deferral approval before proceeding |
When the acceptance-tester reports failures:
brainstormingexecuting-plans and create targeted fix tasksrequesting-code-review on the fixesacceptance-testing after fixes are completeThe fix loop is: executing-plans → requesting-code-review → acceptance-testing. Repeat until all criteria pass.
finishing-a-development-branch while any criterion shows FAILCalled after:
superpowers:requesting-code-review — must complete before acceptance testing beginsCalls:
superpowers:acceptance-tester (subagent, via Task tool)superpowers:executing-plans (on failure, to create fix tasks)superpowers:finishing-a-development-branch (on full pass)If AC document is missing:
superpowers:writing-acceptance-criteria — produce the document firstRequired external skill:
superpowers:playwright-cli — needed for UI interaction criteria; install from https://github.com/microsoft/playwright-cli (see the skills/ directory) if unavailableSee subagent template at: acceptance-testing/acceptance-tester.md
npx claudepluginhub xhyqaq/superpowers-plus --plugin superpowers-plusDrives coding tasks end-to-end from issue intake through implementation, review, deployment, and acceptance verification as a state machine with explicit stop conditions.
Verifies implementation against a spec with evidence-based checks and three independent self-consistency passes. Ensures every requirement is backed by verbatim evidence before merge.
Verifies feature completion by requiring automated tests that prove functionality works. Enforces phase gates and spec alignment before acceptance.