Skip to content

Latest commit

 

History

History
43 lines (30 loc) · 4.36 KB

File metadata and controls

43 lines (30 loc) · 4.36 KB

Agile V Installation Profiles

Repository version: 3.9.x Status: Profile definitions are documentation, not package manifests. Copy each listed skill directory into a skill-discovery directory supported by your agent.

Use the smallest profile that meets the work's risk and lifecycle needs. Add one domain build agent from domains/ when implementation is in scope.

Profile Use when Skill directories
core-minimal Requirements, review, or lifecycle guidance without implementation agile-v-core, requirement-architect, logic-gatekeeper
verified-build Building from an approved baseline with independently designed tests and independent verification agile-v-core, requirement-architect, logic-gatekeeper, build-agent, test-designer, red-team-verifier, agile-v-quality-gates
existing-repo Changing an existing codebase with impact and regression evidence core-minimal + skills/system-understanding-agent, skills/impact-analysis-agent, skills/regression-selection-agent, skills/graph-traceability-agent, skills/diff-evidence-agent + the build/verification skills from verified-build
regulated L3/L4, safety-relevant, sensitive, regulated, or externally assured work verified-build + agile-v-compliance, agile-v-control-matrix, compliance-auditor, validation-agent, safety-engineer, threat-modeler, release-manager, observability-planner; add preview agile-v-aibom and preview agile-v-human-oversight only under an approved local policy
business-preview Evaluating unreleased business and executive-governance contracts agile-v-core, c-suite-foundation, relevant chief-* orchestrators, relevant functional skills (venture-strategist, rd-innovator, gtm-executor, business-operations), optionally c-suite-update
gxp-lean-csa [Draft] Lower-risk, configurable, or well-understood software where approved local procedures permit a lean risk-based approach agile-v-core, agile-v-gxp-qualification, agile-v-control-matrix, agile-v-compliance, requirement-architect, logic-gatekeeper, test-designer, red-team-verifier, validation-agent, compliance-auditor, release-manager
gxp-standard [Draft] Regulated computerized systems requiring explicit DQ/IQ/OQ/PQ packaging gxp-lean-csa + system description, qualification plan, DQ report, IQ/IOQ protocol+report, OQ protocol+report, PQ/intended-use plan+protocol+report, qualification summary, supplier suitability assessment, requalification strategy
gxp-high-assurance [Draft] L4, safety-relevant, high data-integrity risk, externally assured, or critical operational systems gxp-standard + independent quality approval, stronger role separation, signed/integrity-protected records (in the consuming environment), recovery demonstration, witnessed/independently reviewed critical executions, supplier/service-continuity evidence, periodic review + requalification evidence, reference fault-injection/seeded-defect challenge where feasible

business-preview and any individually marked preview skill are not stable operational contracts. Review and baseline their outputs locally before relying on them.

Install

Choose a destination supported by your tool, for example .claude/skills/, .cursor/skills/, .github/skills/, .agents/skills/, or the corresponding user-level directory. From a checkout of this repository:

mkdir -p <skills-dir>
cp -R agile-v-core requirement-architect logic-gatekeeper <skills-dir>/

For a verified Python build, extend that installation:

cp -R build-agent test-designer red-team-verifier agile-v-quality-gates <skills-dir>/
cp -R domains/build-agent-python <skills-dir>/

Replace build-agent-python with build-agent-js, build-agent-nestjs, build-agent-dart, or build-agent-embedded as appropriate. For nested existing-repository skills, copy the individual directories from skills/ into the destination.

Verify Installation

Ask the agent to list or load agile-v-core, then request an ambiguous implementation. Correct behavior is to classify risk, request missing context, and persist requirements before synthesis. Skill discovery and invocation details are tool-specific; installation alone does not prove lifecycle conformance.

Proceed with the Golden Journey for the canonical end-to-end flow.