Use ConvergeQA from Codex, Claude Code, Cowork, OpenClaw, Hermes, or similar AI coworker environments.

---
name: convergeqa-compare-agent
description: Run flexible ConvergeQA Compare workflows. Use when the user asks for ConvergeQA Compare, /convergeqa compare, /compare, to compare a prompt, question, draft, document, source packet, model answer, research direction, code diff, implementation packet, or prepared review packet across a ConvergeQA model panel; when the user wants raw Compare model outputs; when the user asks for a recommended Premium general, Budget general, or saved/custom Coding Review panel; or when the user asks for synthesis. Do not use for Iterate or Critique.
---

# ConvergeQA Compare Agent

Use this skill when the user asks to compare a prompt, question, draft, document, source packet, model answer, research direction, code diff, implementation packet, or prepared review packet across multiple models. Do not use this skill for revision loops or finished-deliverable audits.

## Non-negotiables

- Use the supported Compare due-diligence client when available:

tools/agent_clients/compare_due_diligence_cli.py

- Prefer the client, wrapper, or background worker provided by the project over browser automation.
- Keep credentials in environment variables. Developer API keys use CONVERGEQA_API_KEY and X-API-Key. Org service-account keys use CONVERGEQA_SERVICE_ACCOUNT_KEY and X-Service-Account-Key on approved private agent paths.
- Do not submit PHI, secrets, private credentials, regulated data, or unapproved customer data.
- Recommend a panel before running: Premium general for serious general work, Budget general for quick or low-stakes scans, and a saved or custom Coding Review panel for code or implementation review. Do not treat Coding Review as a pricing tier.
- Respect explicit overrides such as Premium, Budget, Coding Review, no specialists, same prompt, or synthesis off.
- Send the same prompt to the panel by default for ordinary Compare. Use specialist prompts only when the active runner explicitly supports that mode and the user approves the routing.
- When specialist routing is selected or recommended, scan the live visible template library when possible, including built-in and custom templates returned by the template-list API. Recommend the best-fitting 2-4 options before live use. Use General only for generic same-prompt Compare or when no closer template fits.
- Synthesis is on by default for ordinary Compare. Use the no-synthesis flow only when the user explicitly declines synthesis, and keep the selected tier, panel, synthesis model, and credit/billing status visible before live use. Do not surface internal provider billing details to users.
- Return exact model outputs and exact ConvergeQA synthesis before adding any assistant summary. Label any assistant note separately.
- Export or save the packet JSON and bundle ZIP when those artifacts are available.
- Stop before live use when the source is unclear, sensitivity is unresolved, the selected routing is unsupported, or the credit/billing boundary is unclear.

## Command triggers

Treat these as Compare requests:

- ConvergeQA Compare
- /convergeqa compare
- /compare
- compare this with ConvergeQA
- run a Coding Review panel
- review this implementation packet with Compare

## Setup to show before running

Show a concise setup card before the first live call:

- Mode: Compare
- Title or short input label
- Input source: pasted text, local file, URL, or prepared packet
- Template recommendation: selected or recommended template, reason, and alternatives including General or custom template if useful
- Recommended panel: Premium general, Budget general, saved/custom Coding Review, or Coding Review no specialists, with a short reason
- Tier and credit estimate: Budget generally uses 1 credit; Premium or premium-class custom panels generally use 2 credits unless a large/high-token review requires confirmation for more
- Model panel and count
- Prompt routing: same prompt or approved specialist prompts
- Synthesis: on by default unless explicitly declined; show the synthesis model and opt-out path
- Credential kind: Developer API key or org service-account key
- Reference material: provided or none provided
- Additional instructions: audience, focus, source standards, length priority, and things not to expand
- Length or read-time priority when relevant
- Decision control: user-driven by default, or approved agent-driven policy
- Stop condition: complete, failed, unsupported route, configured limit, or user stop
- Credit/billing boundary
- Artifact output folder
- Sensitivity check: confirm no PHI, secrets, private credentials, regulated data, or unapproved customer data

## Workflow

1. Prepare the input outside the chat when possible, especially for long files.
2. Show the setup card and wait for approval if any spend, synthesis, sensitive-source, template, specialist-routing, or model-panel choice is unclear.
3. Run the supported Compare client.
4. Poll until the packet is complete, failed, or stopped by the configured limit.
5. Save the packet JSON and bundle ZIP when available.
6. Read the saved artifacts, not just the terminal summary.
7. Present the exact per-model responses first.
8. Present the exact ConvergeQA synthesis if generated.
9. Then, if useful, add a clearly labeled assistant summary.
10. Close with artifact paths, credit, billing, or usage metadata when reported, and any limits.

## Closeout

Report:

- Compare mode used
- Template used
- Model tier and panel
- Whether synthesis was requested and generated
- Packet and bundle locations
- Credit, billing, or usage metadata when reported
- Any failures, missing model responses, or unverified claims

ConvergeQA records a multi-model due-diligence process. Do not claim it proves the content is true, legal, medical, complete, or production-ready.

---
name: convergeqa-iterate-agent
description: Use ConvergeQA Iterate to improve or revise a draft through recommendation-by-recommendation decisions. Do not use for Compare or Critique.
---

# ConvergeQA Iterate Agent

Use this skill when the user asks to improve, revise, iterate, strengthen, tighten, or converge a draft through recommendation-by-recommendation decisions. Do not use this skill for side-by-side model comparison or final audit only.

## Non-negotiables

- Use the supported Critique/Iterate client when available:

tools/agent_clients/service_account_reviews_cli.py --mode iterate

- Prefer the client, wrapper, or background worker provided by the project over browser automation.
- Keep credentials in environment variables. Developer API keys use CONVERGEQA_API_KEY and X-API-Key. Org service-account keys use CONVERGEQA_SERVICE_ACCOUNT_KEY and X-Service-Account-Key on approved private agent paths.
- Do not submit PHI, secrets, private credentials, regulated data, or unapproved customer data.
- Scan the live visible template library when possible, including built-in and custom templates returned by the template-list API. Recommend 2-4 likely templates from the draft or source material, with the best fit first. Do not hardcode a template without reviewing the content.
- Default to Premium tier and note that Budget is available.
- Default to user-driven decisions. Pause after each recommendation packet.
- Use agent-driven decisions only when the user has approved a written decision policy, authority boundary, stop condition, and round limit before the first live call.
- Render ConvergeQA recommendations exactly before asking for accept or decline decisions. Label any assistant note separately.
- Do not rewrite the draft yourself as a substitute for the ConvergeQA decision flow.
- Export or save the final packet and bundle before cleanup.
- Stop when ConvergeQA reports convergence, the approved round limit is reached, the user says finish, a live-session blocker appears, or the sensitivity/credit boundary is no longer satisfied.

## Setup to show before running

Show a concise setup card before the first live call:

- Mode: Iterate
- Draft title
- Input source: pasted text, local file, URL, or prepared packet
- Template options: 2-4 likely visible built-in or custom options, recommended first, or custom template if no current option fits
- Tier: Premium by default; Budget available
- Model panel
- Reference material: provided or none provided
- Additional instructions: audience, format, tone, source standards, length priority, and things not to expand
- Desired length, read time, or output shape
- Iteration control: one round, continue until approved, or approved round limit
- Decision control: user-driven or approved agent-driven policy
- Agent authority: exact decisions allowed, decisions not allowed, and stop condition when delegated
- Spend or credit boundary
- Artifact output folder
- Sensitivity check: confirm no PHI, secrets, private credentials, regulated data, or unapproved customer data

## Decision grammar

Use simple decisions that can be logged:

- A means accept.
- D means decline.
- A [number] means accept that recommendation.
- D [number] means decline that recommendation.
- A1 and D2 shorthand are acceptable when the number is clear.
- A all means accept all currently shown recommendations.
- D all means decline all currently shown recommendations.
- Notes without an explicit accept are treated as comments or decline instructions, not silent acceptance.
- If the user gives ambiguous instructions, ask for clarification before submitting decisions.

## Workflow

1. Prepare the draft and references outside the chat when possible.
2. Show the setup card and get approval for unclear credit/billing, source, model, or decision-control choices.
3. Start the Iterate run with the supported client.
4. Poll until the recommendation packet is available.
5. Read the packet from saved artifacts when possible.
6. Present each ConvergeQA recommendation exactly, with stable numbering.
7. Collect decisions using the decision grammar.
8. Submit the decision packet.
9. Continue only within the approved iteration control.
10. Finish by exporting or saving the final packet and bundle.

## Closeout

Report:

- Iterate mode used
- Model tier and panel
- Rounds completed
- Decision log summary
- Final packet and bundle locations
- Credit, billing, or usage metadata when reported
- Any recommendations skipped, declined, or left unresolved
- Any stop condition reached, skipped export, or unverified claim

ConvergeQA records the revision decision process. Do not claim the result is certified, legally approved, medically safe, or complete unless an appropriate human reviewer has separately confirmed it.

---
name: convergeqa-critique-agent
description: Use ConvergeQA Critique to audit a finished deliverable without editing the original. Do not use for Compare or Iterate.
---

# ConvergeQA Critique Agent

Use this skill when the user asks to audit, critique, review, red-team, or assess a finished deliverable without editing the original. Do not use this skill for ordinary revision loops or model comparison.

## Non-negotiables

- Use the supported Critique runner or Critique client when available:

tools/agent_clients/agent_review_critique_runner.py
tools/agent_clients/service_account_reviews_cli.py --mode critique

- Prefer the client, wrapper, or background worker provided by the project over browser automation.
- Keep credentials in environment variables. Developer API keys use CONVERGEQA_API_KEY and X-API-Key. Org service-account keys use CONVERGEQA_SERVICE_ACCOUNT_KEY and X-Service-Account-Key on approved private agent paths.
- Do not submit PHI, secrets, private credentials, regulated data, or unapproved customer data.
- Treat the original deliverable as unchanged unless the user separately asks for an Iterate pass.
- Scan the live visible template library when possible, including built-in and custom templates returned by the template-list API. Recommend a template or audit lens from the deliverable before live use. Offer 2-4 likely options, recommended first, or a custom template if no current option fits. For patient-facing health education, surface Healthcare: Patient Education when it appears in the visible template library; do not require the user to already know that label.
- Render the full ConvergeQA audit output exactly before asking for decisions. Do not replace it with structured metadata or an assistant-authored summary.
- Critique normally runs one round, then stops at the decision point. Continue only when the user explicitly asks to go deeper or run another round, provides a Disagree reason that needs adjudication, adds new reference or source constraints, or leaves a serious ambiguity that cannot be closed safely. Because the deliverable cannot be edited during the session, Critique should very rarely go past 2-3 rounds; the server enforces a hard cap of 10 rounds.
- Pause for user-driven decisions unless agent-driven authority was approved with a written policy, authority boundary, stop condition, and round limit before the first live call.
- Export or save the audit packet and bundle before cleanup.
- Stop when the audit is complete, the approved round limit is reached, the user says finish, a live-session blocker appears, the sensitivity/credit boundary is no longer satisfied, or a finding requires clinical judgment, professional source judgment, or a choice among conflicting authorities.

## Setup to show before running

Show a concise setup card before the first live call:

- Mode: Critique
- Deliverable title
- Input source: pasted text, local file, URL, or prepared packet
- Deliverable type: PDF, presentation file, document, page, or other
- Template options: 2-4 likely visible built-in or custom options, recommended first, or custom template if no current option fits
- Tier: Premium by default; Budget available
- Model panel
- Reference material: provided or none provided
- Additional instructions: audience, focus, source standards, maximum length or read-time target, and issues not to expand
- Maximum length or read-time nudge if none was provided
- Decision control: user-driven or approved agent-driven policy
- Agent authority: exact decisions allowed, decisions not allowed, and stop condition when delegated
- Spend or credit boundary
- Artifact output folder
- Sensitivity check: confirm no PHI, secrets, private credentials, regulated data, or unapproved customer data

## Decision checkpoint

After showing the full audit, ask what to do next:

- Agree: accept listed findings as-is.
- Disagree [finding numbers]: record disagreement or mark findings as not accepted.
- Go deeper [finding numbers]: request more analysis on specific findings if the runner supports another round.
- Continue: proceed within the approved policy or round limit.
- Finish: export the audit artifacts and stop.

Treat unlisted findings as agreed only when the user explicitly says so. Otherwise keep decisions literal and logged.

## Workflow

1. Prepare the finished deliverable and references outside the chat when possible.
2. Show the setup card and get approval for unclear credit/billing, source, model, or decision-control choices.
3. Start the Critique run with the supported runner or client.
4. Poll until the audit packet is available.
5. Read the packet from saved artifacts when possible.
6. Present the exact ConvergeQA audit, including synthesis and finding structure.
7. Ask for decisions using the decision checkpoint.
8. Continue only within the approved decision control.
9. Finish by exporting or saving the audit packet and bundle.

## Closeout

Report:

- Critique mode used
- Model tier and panel
- Audit packet and bundle locations
- Decisions recorded
- Credit, billing, or usage metadata when reported
- Any unresolved findings, skipped exports, or unverified claims
- Any stop condition reached or official export limitation

ConvergeQA records an audit process. Do not claim it certifies the deliverable as true, legal, medical, complete, or publication-ready unless an appropriate human reviewer has separately confirmed it.