Verification phase

<!-- sflow-copilot-pause -->

Reuse a verified pause-aware entry supplied in this invocation before any mutation or selection change; otherwise first run singularity-flow session current --for-agent --json once; pause precedes Git. If paused: native Copilot, only /sf-pause off, never resume implicitly. Reuse binding/personalization.replyName; fresh operation checks/consent remain required.

<!-- sflow-output-contract: clarification-and-artifact -->

Output contract: Author configured artifacts from governed inputs and pinned clarification; publish only after readiness. Approved phase inputs are the earlier steps' outputs and settle what they state: never ask what they answer or search elsewhere for them. Summarize artifacts with verified Markdown-preview links; never paste document bodies, code or diffs into Copilot. For code, list touched paths and observed added/deleted/created line counts; unavailable stays unavailable. Use existing evidence, not extra calls solely for display. Summaries, links and opened previews are not human review or approval. For every next action, honor returned commandGuidance in order: Copilot, Shell, VS Code (model-free) using modelFreeCommand; if absent, say "not available for this action". Missing Copilot: "no verified equivalent". Never invent routes. On a refusal read remediationPlan: run its first read-only step once and follow context.strategy; never retry an unchanged condition or edit an earlier phase to clear it. One returned automatic plan-hash action (such as sync) may run once, then re-enter once. When a result's act.state is author, or it routes back to this skill with nothing for a person to answer, edit, review or confirm, that is this turn's own work: do it, never relay it. A file edit that does not match is not a refusal: read the file again and redo the edit. Run a check only when a returned field or command asks for it; never decide from a step's name, kind or route that it applies. A refusal saying a check does not apply to this step is not a blocker: skip it and continue. Act on a result's act first: its state, do routes and any ask. A bounded result names agentOutput.read ranges and omittedFields text files; read all in one response with your file tool. If your host saved a result to a file you cannot open, run that command again with --for-agent --json; never read the host's copy with shell tools.

<!-- sflow-execution-boundary -->

Boundary: reuse entry ready/workId, cwd=repositoryPath; CLI/workItemRoot paths only; never $HOME. No duplicate lookup. Use exact CLI-returned paths. repositoryPath is command cwd, not a prefix for absolute file paths. Never rebase paths or strip .git/worktrees. Issued private/external files are exact exceptions to repository confinement, not wider read/write permission; writes still need authorization. Resolve relative paths only against their declared base. On failure, preserve the path and error; never search or recreate it.

1. Run singularity-flow nextsteps --for-agent --json for any workflow. Stop at recovery, pending publication, approval, completion or cancellation. <phase> is its Boundary phase, whatever it is named.
2. Require Boundary ready and exact phaseAgent; otherwise relay the nextsteps route and stop. Run singularity-flow phase show <phase> --for-agent --json. If policyVerified is false, show policyReason, offer /sf-validate then /sf-recover; stop. If effectiveAuthoringSkill is not /sf-phase (a release step returns /sf-release), relay it and stop. Keep Story context governed.
3. Run singularity-flow phase enter <phase> --compose --for-agent --json; use context.text and retainedExecutionEvidence. Never derive --task from Story text.
4. Read the approved phase inputs and changed source.
5. Map each AC to evidence and qualified @ac:WORK-ID:AC-001 in an executable test. Inspect source-bound @clause:WORK-ID:REQ-001 as a trace witness, not a verdict. Reject bare AC-001; honor reviewed test-only/non-code exemptions.
6. Use bound runs over stale prose; IDE discovery is not execution. Honor runner ownership; no substitute commands. Command success is not criterion acceptance. Source edits need governed repair/fresh evidence.
7. Cover relevant regression, boundaries, failures, security, accessibility/performance.
8. Preparation is not authoring. Run singularity-flow prepare <phase>; use the native file-edit tool to complete the existing artifact with observed evidence and Agent brief: verdict, failures/omissions, risk, release recommendation. Resolve TODOs, preserve managed blocks. Failed/unavailable tests are not passes but a verdict to write; never invent results.
9. Run singularity-flow phase prepublish <phase> --for-agent --json --publish-if-ready (includes draft-check). If unready, run read-only singularity-flow recover <WORK-ID> --phase <phase> --for-agent --json. Correct every agent finding now; route other producers to owner via correction.guidance/correction.skill. Never delete markers blindly, invent facts, use padding, invoke nested models, or overwrite another producer. Follow returned repairLoop.protocol for persistent correction/resume; stop unchanged.
10. Use returned commands.publish when ready unless publishIfReady.published; absent: relay commands.next, stop. Race-time ARTIFACT_AUTHORING_INCOMPLETE: recheck once, retry once if ready, never loop. Never submit or approve.
11. From publishIfReady.review, summarize tests, gaps, publication, verified Markdown-preview links, generation/hashes and reviewBinding. No bodies/code/diffs; summaries/links are not human review or reusable approval consent.
12. Never submit/approve. End with each handoff from step 11: Next in Copilot: /sf-… from its copilotCommand, then Terminal equivalent: singularity-flow … from its command.