Choose what to check.

Continuously expand coverage and investigate bugs until stopped or a limit is reached. Requires two separate human confirmations; durable state and protected live report; no automatic fixes.

Requires two distinct confirmations before starting or resuming. Optional bounded limits, pause/stop controls, durable checkpoints and current-work explanations. The host agent must remain active; an HTML viewer is not an autonomous AI process.

Example: /carbon-forever focus on the current project and recent changes

Native MCP prompt: carbon_forever. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Hunt reproducible bugs: short directed AI-code probes, protected open-ended source/browser/API discovery, then evidence confirmation. Exposed as /carbon-issues; no automatic fixes.

Protect at least half the active investigation budget for stateful, multistep and persona exploration when execution is available. Reproduce supported issues, separate static concerns, and disclose blocked execution.

Example: /carbon-issues focus on the current project and recent changes

Native MCP prompt: carbon_issues. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Explain CARBON's capabilities, analyze the current project, and recommend the best workflows and prioritized next steps.

Example: /carbon-help focus on the current project and recent changes

Native MCP prompt: carbon_help. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Create a safe bundled demo project and run a from-scratch CARBON First Look, with or without representative existing tests.

Example: /carbon-demo web-static

Native MCP prompt: carbon_demo. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Manage private local test/model credentials, browser control, reusable context, and optional CARBON preferences.

Editable local settings with global defaults and project overrides, browser control, privacy choices and promotions preferences. Credentials stay protected; do not put keys in chat.

Example: /carbon-settings show my project settings

Native MCP prompt: carbon_settings. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Connect Jira/Xray or TestRail, read external test and issue evidence, manage non-secret project overrides, and preview or explicitly approve guarded synchronization.

Example: /carbon-integrations focus on the current project and recent changes

Native MCP prompt: carbon_integrations. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Show evidence-qualified CARBON activity, token/model usage, issues found and verified fixed, unit economics, command frequency, leverage signals, and optional user-baselined savings.

Show attributable command frequency, outcomes and available usage. Hide unavailable token/model/cost panels instead of inventing measurements. Savings require an explicit baseline.

Example: /carbon-stats focus on the current project and recent changes

Native MCP prompt: carbon_stats. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Explore a durable screenshot/widget quality map, retain human input, and compare added, missing, changed and restored surfaces across builds.

Explore a compact page map. Select one screenshot, zoom an element, inspect recorded checks and the next three candidates. Notes, priorities and custom cases persist across rescans; removed and restored identities retain history. Journey selection opens recorded playback without auto-start.

Example: /carbon-map focus on the current project and recent changes

Native MCP prompt: carbon_map. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Show the complete project-local CARBON quality record: test inventory, declared automation architecture, frameworks, durable pass/fail history and trends, findings, and human-qualified decisions.

Example: /carbon-full-report focus on the current project and recent changes

Native MCP prompt: carbon_full_report. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Ask or record only the project-specific human decisions that can materially change CARBON confidence, readiness, risk, test selection, or release interpretation.

Separate decision-changing questions for testers, project managers, product managers and developers. Retain answers and comments locally; missing human context qualifies confidence without blocking all assessment.

Example: /carbon-human focus on the current project and recent changes

Native MCP prompt: carbon_human. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON Feedback Loop: local artifact feedback or the complete bundled Remote/ReviewLoop lifecycle with reviewer teams, document channels, fixes, retests and a CARBON-branded privately or publicly self-hosted server.

Included CARBON Feedback Loop supports local review and an explicitly configured, self-hosted private/public server. External invitations, uploads and sharing need approval; ordinary HTML email clients cannot be assumed to support embedded forms.

Example: /carbon-feedback focus on the current project and recent changes

Native MCP prompt: carbon_feedback. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Create prompt-defined tests; locally convert TestRail, Xray, and other legacy suites; run evidence-gated bake-in; repair converted definitions; and visually manage intent CRUD.

Example: /carbon-tests focus on the current project and recent changes

Native MCP prompt: carbon_tests. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Generate prioritized planned CARBON tests from discovered widgets, standard flows, code contracts, recordings, and custom or learned flows.

Example: /carbon-generate focus on the current project and recent changes

Native MCP prompt: carbon_generate. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Generate, extend, review, and optionally execute a large categorized chatbot test suite from current conversation context and project artifacts.

Example: /carbon-chatbot focus on the current project and recent changes

Native MCP prompt: chatbot. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Use the coding agent's local browser for classic CARBON browser workflows with no testers.ai account required.

Example: /carbon-browser focus on the current project and recent changes

Native MCP prompt: carbon_browser. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Use CARBON Plan, Test, Review, Focused, or Learn mode with durable local runs, learning, recording, and playback.

Example: /carbon

Native MCP prompt: carbon. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Continue the prior broad CARBON candidate pool with the next highest-value deferred tests without repeating the initial pass.

Example: /carbon-more focus on the current project and recent changes

Native MCP prompt: carbon_more. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Test only the named feature, flow, API, component, or behavior using optional selected quality domains such as sign in plus accessibility.

Example: /carbon-test focus on the current project and recent changes

Native MCP prompt: carbon_test. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON GEO and AI Search testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required GEO report.

Example: /carbon-geo focus on the current project and recent changes

Native MCP prompt: carbon_geo. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON Agent Readiness testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required Agent Readiness report.

Example: /carbon-agentic focus on the current project and recent changes

Native MCP prompt: carbon_agentic. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON Functionality testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required Functionality report.

Example: /carbon-functionality focus on the current project and recent changes

Native MCP prompt: carbon_functionality. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON User Interface testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required UI report.

Example: /carbon-ui focus on the current project and recent changes

Native MCP prompt: carbon_ui. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON User Experience testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required UX report.

Example: /carbon-ux focus on the current project and recent changes

Native MCP prompt: carbon_ux. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON Usability testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required Usability report.

Example: /carbon-usability focus on the current project and recent changes

Native MCP prompt: carbon_usability. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON Accessibility and WCAG testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required Accessibility, WCAG, and ACR report set.

Example: /carbon-accessibility focus on the current project and recent changes

Native MCP prompt: carbon_accessibility. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON Privacy testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required Privacy report.

Example: /carbon-privacy focus on the current project and recent changes

Native MCP prompt: carbon_privacy. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON Security testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required Security report.

Example: /carbon-security focus on the current project and recent changes

Native MCP prompt: carbon_security. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON Load Testing testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required Load testing report.

Read-only source/configuration assessment first. Capture and replay only approved traffic with synthetic identities and bounded stages, request limits and stop conditions. Traffic generation needs explicit user approval.

Example: /carbon-load focus on the current project and recent changes

Native MCP prompt: carbon_load. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON Stress Readiness and Recovery testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required Stress readiness and recovery report.

Prefer disposable localhost tests. Obtain two distinct confirmations before non-local stress traffic; agree workload, ceilings, stop conditions and recovery.

Example: /carbon-stress focus on the current project and recent changes

Native MCP prompt: carbon_stress. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON Performance testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required Performance report.

Example: /carbon-performance focus on the current project and recent changes

Native MCP prompt: carbon_performance. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON Networking testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required Networking report.

Example: /carbon-networking focus on the current project and recent changes

Native MCP prompt: carbon_networking. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON State Integrity and Round Trips testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required State integrity report.

Example: /carbon-state focus on the current project and recent changes

Native MCP prompt: carbon_state. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON Reliability and Resilience testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required Reliability report.

Example: /carbon-reliability focus on the current project and recent changes

Native MCP prompt: carbon_reliability. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON Compatibility testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required Compatibility report.

Example: /carbon-compatibility focus on the current project and recent changes

Native MCP prompt: carbon_compatibility. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON Content testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required Content report.

Example: /carbon-content focus on the current project and recent changes

Native MCP prompt: carbon_content. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON Data Integrity testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required Data integrity report.

Example: /carbon-data focus on the current project and recent changes

Native MCP prompt: carbon_data. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON API testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required API report.

Example: /carbon-api focus on the current project and recent changes

Native MCP prompt: carbon_api. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON Internationalization Readiness and Completion testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required INTL readiness and completion report.

Example: /carbon-intl focus on the current project and recent changes

Native MCP prompt: carbon_intl. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run CARBON Localization and Internationalization testing against a named feature, reachable target, API, source, screenshot, or current project and generate the required Localization report.

Example: /carbon-localization focus on the current project and recent changes

Native MCP prompt: carbon_localization. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run a bounded durable test-fix-retest validation loop.

Example: /carbon-auto focus on the current project and recent changes

Native MCP prompt: carbon_auto. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Create isolated copies, inject controlled faults and error conditions, verify recovery, and prove the originals remained unchanged.

Example: /carbon-failure focus on the current project and recent changes

Native MCP prompt: failure. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Upgrade existing tests with sparse model-powered screenshot quality checkpoints or deterministic local validation.

Example: /carbon-ai-upgrade focus on the current project and recent changes

Native MCP prompt: ai_upgrade. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Run an intent-first confidence investigation loop: prioritize consequential questions, investigate and challenge, reassess evidence, then reprioritize within budget with a live decision audit.

Orient → prioritize → investigate → challenge → assess → reprioritize. The default controller budget is 20 minutes and six investigations. The live audit shows actions, concise rationale and evidence, not hidden reasoning. Reassess after new evidence; humans own release authority.

Example: /carbon-confidence focus on the current project and recent changes

Native MCP prompt: carbon_confidence. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Initialize a durable confidence-engineering workspace.

Example: /carbon-confidence-init focus on the current project and recent changes

Native MCP prompt: carbon_confidence_init. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Map AI-system risks, claims, and decision-changing evidence.

Example: /carbon-confidence-plan focus on the current project and recent changes

Native MCP prompt: carbon_confidence_plan. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Design, run, and analyze decision-relevant AI evaluations.

Example: /carbon-eval focus on the current project and recent changes

Native MCP prompt: carbon_eval. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Independently challenge AI changes and their evidence.

Example: /carbon-review focus on the current project and recent changes

Native MCP prompt: carbon_review. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Decide whether an AI release should ship, canary, hold, or roll back.

Example: /carbon-release focus on the current project and recent changes

Native MCP prompt: carbon_release. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Contain an AI incident and preserve durable regression evidence.

Example: /carbon-incident focus on the current project and recent changes

Native MCP prompt: carbon_incident. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Challenge evaluation populations, cases, and oracles.

Example: /carbon-eval-design-review focus on the current project and recent changes

Native MCP prompt: carbon_eval_design_review. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Independently synthesize release evidence and uncertainty.

Example: /carbon-release-review focus on the current project and recent changes

Native MCP prompt: carbon_release_review. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Review AI trust boundaries, data, tools, and side effects.

Example: /carbon-security-review focus on the current project and recent changes

Native MCP prompt: carbon_security_review. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Find overclaims, weak evidence, and confidence theater.

Example: /carbon-skeptical-review focus on the current project and recent changes

Native MCP prompt: carbon_skeptical_review. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.

Review AI evaluation statistics and supported inference.

Example: /carbon-statistical-review focus on the current project and recent changes

Native MCP prompt: carbon_statistical_review. Use your host’s skill picker or installed alias; command prefixes vary by host. Read the applicable safety boundary before execution.