AgentStack
SKILL verified Apache-2.0 Self-run

Participant Web Operator

skill-jasontang-ai-acp-participant-web-operator · by jasontang-ai

Use this skill to operate and QA the Relay web app for participant submission, digest reading, baseline thread reading, response flow, and operator review.

No reviews yet
0 installs
2 views
0.0% view→install

Install

$ agentstack add skill-jasontang-ai-acp-participant-web-operator

✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

Are you the author of Participant Web Operator? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

What this skill is for

Use this skill when ACP must be verified through the rendered Relay web surface rather than only CLI or API outputs.

When to use it

  • Participant submission workflow QA.
  • Digest or baseline thread reading workflow QA.
  • Operator UI inspection of Engine V2 issue maps, critics, and release state.
  • Browser/server mismatch investigation.

When not to use it

  • Do not use it for protocol logic that can be checked directly through CLI or schema.
  • Do not use it to bypass server lifecycle transitions.
  • Do not use it when no web surface is involved.

Inputs expected

  • Local app URL
  • Cycle id when targeting a specific workflow
  • Participant id when inspecting participant views
  • Expected server-side state

Preflight / prerequisites

  1. Confirm server readiness.
  2. Run scripts/surface-preflight.mjs --base-url when state is uncertain.
  3. Know whether the task is observation-only or action-authorized.
  4. Keep screenshots or notes tied to the cycle id.

Workflow

  1. Inspect server readiness and participant view payloads.
  2. Open the app in the browser.
  3. Execute the specific participant or operator path.
  4. Compare rendered state against server contract.
  5. Report any mismatch with URL, cycle id, participant id, and expected server state.

Decision rules / judgment criteria

  • Server contract beats visual assumption.
  • UI polish is not success if lifecycle state is wrong.
  • Do not submit participant content unless the user asked for that action.
  • A browser-only observation is not enough for claim-bearing evidence.

Escalation rules

Escalate when the browser and API disagree, when a participant can see the wrong condition surface, or when a UI path permits lifecycle bypass.

Available scripts

  • scripts/surface-preflight.mjs --base-url [--cycle-id ] [--participant-id ]

Outputs

  • Readiness snapshot
  • Browser workflow observations
  • Server/browser mismatch report
  • Operator QA notes

Failure handling

If the browser surface fails but the server is healthy, preserve both observations. If the server contract fails, stop browser QA and fix the backend surface first.

Trust / safety notes

Participant-facing UI can alter trust in ACP. Treat condition leakage, stale digests, and misleading status labels as serious review findings.

Composition notes

Use in surface-preflight, pilot-cycle-review, and Foresight demo rehearsals.

Examples to inspect next

Read examples/casebook.md for happy path, failure path, ambiguity, and anti-pattern examples.

Evaluation hooks

Run scripts/surface-preflight.mjs and npm run skills:audit.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet — be the first.

Versions

  • v0.1.0 Imported from the upstream source.