Why ChimpHands

In brief: ChimpHands gives teams native QA cloud agents with isolated spend, workflow-level audit trails, one-click async orchestration from TestChimp insights, and budget-aware execution via cost-optimized models.

Teams already run AI agents for development. QA workloads have different economics, risk profile, and observability needs. ChimpHands is purpose-built for QA orchestration on your CI, not as a generic coding-agent replacement.

Cost isolation for QA workloads

ChimpHands metered usage is separate from your Cursor, Claude, or other dev-agent subscriptions. QA automations, nightly upkeep, and plan authoring do not consume the tokens you reserved for feature implementation.

That isolation matters because QA workflows are often:

• Higher volume — every failed batch, ready story, or release gate can spawn a run.
• More tool-heavy — SmartTests, MCP calls, and repo-wide searches add up quickly on premium dev models.
• Lower tolerance for surprise bills — you want predictable spend tied to QA outcomes, not mixed into dev tooling invoices.

ChimpHands credits are tracked per organization with included allowance per billing cycle and transparent usage in the product. Workflow executions retain the full prompt, session link, and status timeline for auditability at the workflow level—who triggered what, when, and what the agent did.

Save dev-agent subscriptions for dev work

Your Claude Code, Cursor, or Copilot seats are best spent on implementation and review, where you want the fastest frontier models and tight IDE integration.

ChimpHands handles the long tail of QA execution:

• Fix failing SmartTests after CI
• Run composite run-qa on a release branch
• Instrument TrueCoverage and close coverage gaps
• Upkeep and cleanup across the suite

You get a dedicated QA agent surface without competing for the same subscription quotas you rely on during the day.

One-click async QA orchestration

TestChimp already tracks issues, plans, test batches, releases, TrueCoverage gaps, and requirement quality. Automations connect those signals to catalog workflows.

With ChimpHands as the invocation strategy, you configure automations once and let the platform:

1. Match events under your conditions and aggregation policy.
2. Optionally wait for human approval.
3. Dispatch the workflow to ChimpHands on your CI with the rendered /testchimp prompt and workflow-execution-id.

No labelled GitHub issues to triage, no third-party agent API keys, no custom webhook receivers—native dispatch with session and execution links in TestChimp.

Platform-native context in chat

Generic cloud agents make you paste issue numbers, hunt for batch ids, or build custom MCP glue to reach TestChimp data. ChimpHands chat includes:

• Entity picker — stories, scenarios, issues, tests, releases, and automation batches as removable composer tags.
• Serialized refs — stable ordinal ids and platform identifiers in the prompt, not prose descriptions.
• Plans / Issues handoff — open ChimpHands with entity (and optional line-range) context already attached.

Combined with repo line ranges, file upload, and collaborative editing in the diff pane, you get IDE-agent ergonomics plus QA-platform context—without committing mcp.json or configuring MCP on a third-party host. See ChimpHands introduction → Chat with platform and repo context.

Cheap tokenomics

ChimpHands routes LLM traffic through cost-optimized models in the GPT 5.6 range rather than always calling the most expensive frontier tier. QA workflows are playbook-driven: much of the work is structured tool use (MCP, git, test runners) where a capable but efficient model is the right tradeoff.

TestChimp operates the routing and metering so you are not managing model API keys per repo for QA runs—the platform optimizes for low marginal cost per workflow execution.

Budget-aware QA protection

You can set how much you want to spend on ChimpHands-driven QA in a billing cycle. The platform enforces budget gates before starting metered work and surfaces remaining credits in the ChimpHands UI.

Within the budget, ChimpHands aims to deliver the best feasible protection:

• Prioritize high-signal automations (severity, release gates, failed batches).
• Run upkeep and gap-closing workflows that leverage platform insights.
• Pause or fail closed when credits are exhausted rather than silently running up an unbounded bill.

Raising the budget unlocks more async coverage; tightening it keeps QA agent spend predictable while you tune which automations fire.

ChimpHands vs bring-your-own cloud agents

	ChimpHands	Via GitHub Issue / Webhook
Setup	GitHub App + one workflow file	Skill, MCP, secrets, per-agent wiring
Skill + CLI on runner	Auto-installed each job	You maintain skill + mcp.json per host
Platform context in chat	Entity picker + refs built in	Manual ids or custom integration
Runs on	Your GitHub Actions runner	Your agent host (varies)
Session continuity	Same session async ↔ chat	Usually separate systems
Cost model	TestChimp ChimpHands credits	Your agent provider's billing
Best for	Default QA orchestration	Custom agent stacks or non-GitHub runners

For most teams, ChimpHands is the default. Use GitHub Issue or Webhook only when you must keep an existing agent product as the executor.

Next steps

• Meet ChimpHands (announcement)
• ChimpHands introduction — Architecture and human-in-the-loop
• Best practices — How to get the most from ChimpHands
• Create an automation — Wire events to ChimpHands
• Typical automation setups — Ready-made recipes