Plugin Claude Code
VV Claude Code Harness plugin marketplace.
Plugin Claude Code
VV Claude Code Harness plugin marketplace.
Plugin Claude Code
Multi-session continuity, worktree-isolated parallel workflows, and mechanical quality enforcement for Claude Code.
Hook Claude Code
Runs before the agent uses a tool for Edit, Write, MultiEdit and Bash tool calls, executing enforce-scope.sh, verify-git-identity.sh and commit-gate.sh (4 commands). From oeftimie/vv-claude-harness.
Hook Claude Code
Runs after a tool call finishes for Edit, Write and MultiEdit tool calls, running an inline shell check. From oeftimie/vv-claude-harness.
Hook Claude Code
Runs when a task is marked completed, executing verify-task-quality.sh. From oeftimie/vv-claude-harness.
Settings file Claude Code
Agent settings declaring 3 hook events (PreToolUse, PostToolUse, TaskCompleted) and 9 allowed tools.
Instructions file CodexOpenCode
Instructions for oeftimie/vv-claude-harness, covering vv claude code harness — distribution repository, what this repo contains, key distinction and working on this repo.
Instructions file
Instructions for oeftimie/vv-claude-harness, covering claude code, harness, harness prep: risk/lane self-classification and path resolution and hook testing.
Agent
Author-blind behavior-test writer (F017/OVI-65). Derives conformance tests from a feature's verified description and acceptance criteria alone -- never from the implementation diff, the implementer's completion message, or any test file the implementer wrote. Adapted from agent-os's conformance-test-writer pattern…
Agent
Harness workflow agent that implements exactly one assigned feature within its assigned scope using strict TDD. Spawn worktree-isolated via the harness-continue workflow with per-feature specifics (feature ID, scope, deliverable, task ID) in the prompt.
Agent
Harness workflow agent that owns one architectural layer (e.g. API handlers, data layer) other agents depend on. Builds to the shared interfaces named in its spawn prompt. Spawn via the harness-continue workflow with layer, scope, and interface partners in the prompt.
Agent
Harness research agent. Answers a scoped research question in one focused pass and writes findings to the file the lead names. Never implements or modifies code. Spawn via the harness-continue workflow with the question and output file in the prompt.
Agent
Integrity check on the spec gate's human loop. Runs when a human answers or amends a spec that spec-verification questioned: re-runs every SV check on the amended text, hosts the anti-sycophancy and anti-capitulation clauses, and refuses to advance on pressure alone. Spawned read-only by the harness-issue-prep skill…
Agent
Harness review agent. Senior review of completed features for correctness, scope adherence, test quality, and the 95% coverage gate. Cannot edit files by construction (no Edit/Write tools); Bash is limited to test runs and read-only git inspection (git diff, git show) by instruction. Reports findings to the lead in…
Agent
Spec gate of the vv-harness pipeline. Proves a specification (a Linear issue or a proposed features.json feature set) is complete, testable, and unambiguous before any implementation starts, or returns precise numbered questions to a human. Spawned read-only by /harness-init Step 5.1 and the harness-issue-prep skill…
Hook
Runs when a session starts on startup, resume, clear and compact, executing session-start.sh. From oeftimie/vv-claude-harness.
Hook
Runs when a session ends, executing session-end.sh. From oeftimie/vv-claude-harness.
Skill Claude CodeCodex
Continue working on a harness-managed project (vv-harness plugin). Orients to current state, picks single-session or workflow mode, and guides implementation with TDD, quality gate hooks, and compaction-aware context management. Use at the start of any session on a harness project.
Skill Claude CodeCodex
Report-first, idempotent instance health check for a harness-managed project. Verifies python3/git presence, the hook set and its executability, .claude/settings.json wiring, .gitignore rules, .harness/ file validity, version drift against the current plugin, whether a passing/in-progress feature's testfile actually…
Skill Claude CodeCodex
Run an observation-first improvement loop on a harness-managed project -- record a job contract, observe the baseline, locate the earliest failed handoff, classify the gap to its smallest vv-native owner, implement one intervention, verify at the claim boundary, rerun fresh, then retain/revise/remove. Use when a user…
Skill Claude CodeCodex
Initialize a new project with the Long-Running Agent Harness (vv-harness plugin). Sets up feature tracking, git identity capture, context summary, build hooks, quality gate hooks, and an optional workflow size guideline. Use when starting a new multi-session project.
Skill Claude CodeCodex
Open a failed or parked piece of work in a live repair session. Input is a features.json feature id (harness project) or a Linear issue key parked by the external runner. Loads the failure context, checks out the branch, iterates the fix with the human, and exits by resuming the runner, routing back through…
Skill Claude CodeCodex
Interactively verify a spec until it is buildable, then mark it ready for implementation. Sources: a Linear issue (via the Linear MCP), a pasted spec, or an existing features.json feature. On PASS it normalizes the spec, records verification (spec field locally; a signed readiness stamp plus label on Linear), and can…