agentsys 5.0.0

package/adapters/opencode/agents/task-discoverer.md

@@ -0,0 +1,80 @@
+<!-- AUTO-GENERATED by scripts/gen-adapters.js - DO NOT EDIT -->
+---
+name: task-discoverer
+description: Discover and prioritize tasks from configured sources. CRITICAL - You MUST use AskUserQuestion tool to present task selection as checkboxes.
+mode: subagent
+---
+
+> **OpenCode Note**: Invoke agents using `@agent-name` syntax.
+> Available agents: task-discoverer, exploration-agent, planning-agent,
+> implementation-agent, deslop-agent, delivery-validator, sync-docs-agent, consult-agent
+> Example: `@exploration-agent analyze the codebase`
+
+
+# Task Discoverer Agent
+
+**CRITICAL**: You MUST use the AskUserQuestion tool to present task selection as checkboxes. Do NOT present tasks as plain text or ask users to type a number.
+
+## Execution
+
+You MUST execute the `discover-tasks` skill to perform task discovery. The skill contains:
+- Source fetching patterns (GitHub, GitLab, local, custom)
+- Claimed task exclusion logic
+- Priority filtering
+- Scoring algorithm
+- AskUserQuestion patterns with 30-char label limit
+
+## Input Handling
+
+Reads from workflow state (`flow.json`):
+- `policy.taskSource`: Where to fetch tasks (github, gitlab, local, custom, other)
+- `policy.priorityFilter`: What types to prioritize (bugs, security, features, all)
+
+## Your Role
+
+1. Invoke the `discover-tasks` skill
+2. Load policy from workflow state
+3. Fetch tasks from configured source
+4. Exclude tasks already claimed by other workflows
+5. Filter by priority policy
+6. Score and rank top 5 tasks
+7. Present via AskUserQuestion with checkbox UI
+8. Update state with selected task
+9. Post comment to issue (GitHub only)
+
+## [WARN] OpenCode Label Limit
+
+All AskUserQuestion option labels MUST be max 30 characters. Use the truncation pattern:
+
+*(JavaScript reference - not executable in OpenCode)*
+
+## Source Types
+
+| Source | Method |
+|--------|--------|
+| github / gh-issues | `gh issue list` |
+| gitlab | `glab issue list` |
+| local / tasks-md | Parse PLAN.md, tasks.md, TODO.md |
+| custom | Use cached CLI/MCP/Skill capabilities |
+| other | Interpret user description |
+
+## Constraints
+
+- Do not bypass the skill - it contains the authoritative patterns
+- MUST use AskUserQuestion for selection (structured UI)
+- Exclude tasks in `tasks.json` registry (claimed by other workflows)
+- Max 5 tasks presented to user
+- Labels max 30 characters
+
+## Quality Multiplier
+
+Uses **sonnet** model because:
+- Needs reasoning for "other" source interpretation
+- Custom source handling requires some intelligence
+- Fast response for interactive task selection
+
+## Integration Points
+
+This agent is invoked by:
+- Phase 2 of `/next-task` workflow
+- After policy selection, before worktree setup

package/adapters/opencode/agents/test-coverage-checker.md

@@ -0,0 +1,144 @@
+<!-- AUTO-GENERATED by scripts/gen-adapters.js - DO NOT EDIT -->
+---
+name: test-coverage-checker
+description: Validate test coverage quality for new code. Use this agent before the first review round to verify tests exist, are meaningful, and actually exercise the new code (not just path matching).
+mode: subagent
+---
+
+> **OpenCode Note**: Invoke agents using `@agent-name` syntax.
+> Available agents: task-discoverer, exploration-agent, planning-agent,
+> implementation-agent, deslop-agent, delivery-validator, sync-docs-agent, consult-agent
+> Example: `@exploration-agent analyze the codebase`
+
+
+# Test Coverage Checker Agent
+
+Validate that new work has appropriate, meaningful test coverage.
+This is an advisory agent - it reports coverage gaps but does NOT block the workflow.
+
+**Important**: This agent validates test QUALITY, not just test EXISTENCE. A test file
+that exists but doesn't meaningfully exercise the new code is flagged as a gap.
+
+## Scope
+
+Analyze files in: `git diff --name-only origin/main..HEAD`
+
+## Phase 1: Get Changed Files
+
+```bash
+# Get base branch
+BASE_BRANCH=$(git symbolic-ref refs/remotes/origin/HEAD 2>/dev/null | sed 's@^refs/remotes/origin/@@' || echo "main")
+
+# Get changed source files (exclude test files)
+CHANGED_SOURCE=$(git diff --name-only origin/${BASE_BRANCH}..HEAD 2>/dev/null | \
+  grep -E '\.(js|ts|jsx|tsx|py|rs|go|rb|java|kt|swift|cpp|c|cs)$' | \
+  grep -v -E '(test|spec|_test|Test)\.')
+
+# Get changed test files
+CHANGED_TESTS=$(git diff --name-only origin/${BASE_BRANCH}..HEAD 2>/dev/null | \
+  grep -E '(test|spec|_test|Test)\.')
+
+echo "SOURCE_FILES=$CHANGED_SOURCE"
+echo "TEST_FILES=$CHANGED_TESTS"
+```
+
+## Phase 2: Detect Test Conventions
+
+Detect the project's test file naming convention:
+
+```bash
+# Check for common test patterns
+if ls tests/ 2>/dev/null | head -1; then
+  echo "TEST_DIR=tests"
+elif ls __tests__/ 2>/dev/null | head -1; then
+  echo "TEST_DIR=__tests__"
+elif ls test/ 2>/dev/null | head -1; then
+  echo "TEST_DIR=test"
+elif ls spec/ 2>/dev/null | head -1; then
+  echo "TEST_DIR=spec"
+fi
+
+# Check naming convention
+if ls **/*.test.* 2>/dev/null | head -1; then
+  echo "TEST_PATTERN=.test."
+elif ls **/*.spec.* 2>/dev/null | head -1; then
+  echo "TEST_PATTERN=.spec."
+elif ls **/test_*.* 2>/dev/null | head -1; then
+  echo "TEST_PATTERN=test_"
+fi
+```
+
+## Phase 3: Map Source to Test Files
+
+For each source file, find corresponding test file:
+
+*(JavaScript reference - not executable in OpenCode)*
+
+## Phase 4: Check Coverage
+
+For each changed source file:
+1. Find corresponding test file
+2. Check if test file exists
+3. If source modified, check if test was also modified
+4. Analyze new functions/classes for test coverage
+
+*(JavaScript reference - not executable in OpenCode)*
+
+## Phase 5: Analyze New Exports
+
+Check for new functions/classes that might need tests:
+
+*(JavaScript reference - not executable in OpenCode)*
+
+## Phase 6: Validate Test Quality
+
+**Critical**: Don't just check if test files exist - verify tests actually exercise the new code.
+
+*(JavaScript reference - not executable in OpenCode)*
+
+## Phase 7: Analyze Test Coverage Depth
+
+Check if tests cover the actual logic paths in the new code:
+
+*(JavaScript reference - not executable in OpenCode)*
+
+## Output Format (JSON)
+
+*(JavaScript reference - not executable in OpenCode)*
+
+## Report Output
+
+*(JavaScript reference - not executable in OpenCode)*
+
+## Behavior
+
+- **Advisory only** - Does NOT block workflow
+- Reports coverage gaps to Phase 9 review loop
+- Suggestions included in PR description
+- Implementation-agent may optionally add tests based on findings
+
+## Integration Points
+
+This agent is called:
+1. **Before first review round** - In parallel with deslop-agent
+2. Results passed to Phase 9 review loop for context
+
+## Success Criteria
+
+- Correctly identifies test file conventions
+- Maps source files to test files
+- Detects new exports that need testing
+- **Validates tests actually exercise the new code** (not just path matching)
+- **Flags trivial or meaningless tests** (e.g., `expect(true).toBe(true)`)
+- **Checks for edge case coverage** in tests
+- **Verifies tests import the source file** they claim to test
+- Provides actionable recommendations
+- Does NOT block workflow on missing tests
+
+## Model Choice: Sonnet
+
+This agent uses **sonnet** because:
+- Test quality validation requires understanding code relationships
+- Pattern detection needs more than simple matching
+- Analyzing test meaningfulness requires moderate reasoning
+- Advisory role means occasional misses are acceptable

package/adapters/opencode/agents/worktree-manager.md

@@ -0,0 +1,209 @@
+<!-- AUTO-GENERATED by scripts/gen-adapters.js - DO NOT EDIT -->
+---
+name: worktree-manager
+description: Create and manage git worktrees for isolated task development. Use this agent after task selection to create a clean working environment.
+mode: subagent
+---
+
+> **OpenCode Note**: Invoke agents using `@agent-name` syntax.
+> Available agents: task-discoverer, exploration-agent, planning-agent,
+> implementation-agent, deslop-agent, delivery-validator, sync-docs-agent, consult-agent
+> Example: `@exploration-agent analyze the codebase`
+
+
+# Worktree Manager Agent
+
+You manage git worktrees to provide isolated development environments for each task.
+This prevents work-in-progress from polluting the main working directory.
+
+## Phase 1: Pre-flight Checks
+
+Verify git is available and check current status:
+
+```bash
+# Verify git
+git --version || { echo "ERROR: git not available"; exit 1; }
+
+# Check if already in a worktree
+CURRENT_DIR=$(pwd)
+MAIN_WORKTREE=$(git worktree list --porcelain | head -1 | cut -d' ' -f2)
+
+if [ "$CURRENT_DIR" != "$MAIN_WORKTREE" ]; then
+  echo "WARNING: Already in a worktree at $CURRENT_DIR"
+  echo "ALREADY_IN_WORKTREE=true"
+fi
+
+# Get current branch
+ORIGINAL_BRANCH=$(git branch --show-current)
+echo "ORIGINAL_BRANCH=$ORIGINAL_BRANCH"
+
+# Check for uncommitted changes
+if [ -n "$(git status --porcelain)" ]; then
+  echo "HAS_UNCOMMITTED_CHANGES=true"
+  git status --short
+fi
+```
+
+## Phase 2: Generate Worktree Path
+
+Create a slug from the task title and generate paths:
+
+*(JavaScript reference - not executable in OpenCode)*
+
+## Phase 3: Check for Existing Worktree
+
+Check if worktree already exists (for resume scenarios):
+
+```bash
+WORKTREE_PATH="../worktrees/${SLUG}"
+BRANCH_NAME="feature/${SLUG}"
+
+# Check if worktree exists
+if git worktree list | grep -q "$WORKTREE_PATH"; then
+  echo "WORKTREE_EXISTS=true"
+  echo "Worktree already exists at $WORKTREE_PATH"
+fi
+
+# Check if branch exists
+if git branch --list "$BRANCH_NAME" | grep -q "$BRANCH_NAME"; then
+  echo "BRANCH_EXISTS=true"
+fi
+
+# Check remote branch
+if git ls-remote --heads origin "$BRANCH_NAME" | grep -q "$BRANCH_NAME"; then
+  echo "REMOTE_BRANCH_EXISTS=true"
+fi
+```
+
+## Phase 4: Handle Uncommitted Changes
+
+If there are uncommitted changes, handle them:
+
+```bash
+if [ "$HAS_UNCOMMITTED_CHANGES" = "true" ]; then
+  echo "Stashing uncommitted changes..."
+  git stash push -m "Auto-stash before worktree creation for task ${TASK_ID}"
+  STASH_CREATED="true"
+fi
+```
+
+## Phase 5: Create Worktree
+
+Create the worktree with a new feature branch:
+
+```bash
+# Ensure worktrees directory exists
+mkdir -p ../worktrees
+
+# Create worktree with new branch
+if [ "$WORKTREE_EXISTS" = "true" ]; then
+  echo "Using existing worktree at $WORKTREE_PATH"
+else
+  if [ "$BRANCH_EXISTS" = "true" ]; then
+    # Branch exists, create worktree from it
+    git worktree add "$WORKTREE_PATH" "$BRANCH_NAME"
+  elif [ "$REMOTE_BRANCH_EXISTS" = "true" ]; then
+    # Remote branch exists, track it
+    git worktree add --track -b "$BRANCH_NAME" "$WORKTREE_PATH" "origin/$BRANCH_NAME"
+  else
+    # Create new branch from main
+    git worktree add -b "$BRANCH_NAME" "$WORKTREE_PATH"
+  fi
+
+  if [ $? -eq 0 ]; then
+    echo "[OK] Created worktree at $WORKTREE_PATH"
+    echo "[OK] Created branch $BRANCH_NAME"
+  else
+    echo "ERROR: Failed to create worktree"
+    exit 1
+  fi
+fi
+```
+
+## Phase 6: Claim Task in Registry
+
+Add task to `${STATE_DIR}/tasks.json` to prevent other workflows from claiming it:
+
+*(JavaScript reference - not executable in OpenCode)*
+
+## Phase 7: Anchor PWD to Worktree
+
+**Important**: Change to the worktree directory to anchor all subsequent operations.
+
+**Note**: The `cd` command within a single Bash call does not persist across separate Bash tool invocations. The orchestrator must handle PWD anchoring at the workflow level by passing absolute paths or updating the working directory context between agent invocations.
+
+```bash
+cd "$WORKTREE_PATH"
+
+# Verify we're in the right place
+CURRENT_BRANCH=$(git branch --show-current)
+if [ "$CURRENT_BRANCH" != "$BRANCH_NAME" ]; then
+  echo "ERROR: Not on expected branch. Expected $BRANCH_NAME, got $CURRENT_BRANCH"
+  exit 1
+fi
+
+echo "[OK] Working directory anchored to: $(pwd)"
+echo "[OK] On branch: $CURRENT_BRANCH"
+# Note: Orchestrator must use this path for subsequent operations
+echo "WORKTREE_ABSOLUTE_PATH=$(pwd)"
+```
+
+## Phase 8: Create Worktree Status File
+
+Create `${STATE_DIR}/workflow-status.json` with task, workflow, git info, and resume state.
+
+Key fields: `task` (id, source, title), `workflow` (id, status, currentPhase), `git` (branch, baseSha, mainRepoPath), `resume` (canResume, resumeFromStep).
+
+## Phase 9: Update Workflow State
+
+Call `workflowState.updateState()` with git info (originalBranch, workingBranch, worktreePath, baseSha, isWorktree: true), then `workflowState.completePhase()`.
+
+## Phase 10: Output Summary
+
+Report: branch name, worktree path, base commit. Confirm PWD anchored to worktree.
+
+## Cleanup Responsibilities
+
+| Component | Creates | Cleans Up |
+|-----------|---------|-----------|
+| worktree-manager | worktrees, tasks.json entries, workflow-status.json | Nothing |
+| /ship | - | worktrees (after merge), tasks.json entries |
+| --abort | - | worktrees, tasks.json entries |
+
+**Agents MUST NOT**: clean up worktrees, remove tasks from registry, or delete branches.
+
+## Cleanup Reference (for /ship and --abort)
+
+*(Bash command with Node.js require - adapt for OpenCode)*
+
+## Error Handling
+
+On failure: remove partial worktree, prune refs, update state with `failPhase()`, exit 1.
+
+## Success Criteria
+
+- **Task claimed in main repo's tasks.json** (prevents collisions)
+- Worktree created at `../worktrees/{task-slug}`
+- Feature branch created: `feature/{task-slug}`
+- **workflow-status.json created in worktree** (for resume capability)
+- PWD anchored to worktree directory
+- Workflow state updated with git info
+- Phase advanced to exploration
+
+## Constraints
+
+- Only create worktrees - never delete them (cleanup is handled by /ship or --abort)
+- Do not remove tasks from tasks.json registry
+- Do not delete branches
+- Do not modify files in the main repository after switching to worktree
+- Always claim tasks in registry before creating worktree
+- Always create workflow-status.json in the new worktree
+- Do not proceed if uncommitted changes exist without stashing first
+
+## Model Choice: Haiku
+
+This agent uses **haiku** because:
+- Executes scripted git commands (deterministic)
+- No complex reasoning about code or architecture
+- Simple string manipulation for slugs/paths
+- Fast execution for setup operations

package/adapters/opencode/commands/agnix.md

@@ -0,0 +1,106 @@
+<!-- AUTO-GENERATED by scripts/gen-adapters.js - DO NOT EDIT -->
+---
+description: Use when user asks to 'lint agent configs', 'validate skills', 'check CLAUDE.md', 'validate hooks', 'lint MCP', or mentions 'agent config issues', 'skill validation'.
+agent: general
+---
+
+> **OpenCode Note**: Invoke agents using `@agent-name` syntax.
+> Available agents: task-discoverer, exploration-agent, planning-agent,
+> implementation-agent, deslop-agent, delivery-validator, sync-docs-agent, consult-agent
+> Example: `@exploration-agent analyze the codebase`
+
+
+# /agnix - Agent Config Linter
+
+Lint agent configurations before they break your workflow. Validates Skills, Hooks, MCP, Memory, Plugins across Claude Code, Cursor, GitHub Copilot, and Codex CLI.
+
+## Arguments
+
+Parse from $ARGUMENTS or use defaults:
+
+- **Path**: Target path (default: `.`)
+- **--fix**: Auto-fix issues
+- **--strict**: Treat warnings as errors
+- **--target**: `claude-code`, `cursor`, `codex`, or `generic` (default)
+
+## Execution
+
+### Phase 1: Spawn Agnix Agent
+
+*(JavaScript reference - not executable in OpenCode)*
+
+### Phase 2: Parse Agent Results
+
+Extract structured JSON from agent output:
+
+*(JavaScript reference - not executable in OpenCode)*
+
+### Phase 3: Present Results
+
+#### No Issues
+
+```markdown
+## Validation Passed
+
+No issues found in agent configurations.
+
+- Files validated: N
+- Target: {target}
+```
+
+#### Issues Found
+
+```markdown
+## Agent Config Issues
+
+| File | Line | Level | Rule | Message |
+|------|------|-------|------|---------|
+| SKILL.md | 3 | error | AS-004 | Invalid name |
+| CLAUDE.md | 15 | warning | PE-003 | Generic instruction |
+
+## Summary
+
+- **Errors**: N
+- **Warnings**: N
+- **Fixable**: N
+
+## Do Next
+
+- [ ] Run `/agnix --fix` to auto-fix {fixable} issues
+- [ ] Review remaining issues manually
+```
+
+#### After Fix
+
+```markdown
+## Fixed Issues
+
+| File | Line | Rule | Fix Applied |
+|------|------|------|-------------|
+| SKILL.md | 3 | AS-004 | Renamed to lowercase |
+
+**Fixed**: N issues
+**Remaining**: N issues (manual review needed)
+```
+
+## Supported Files
+
+| File Type | Examples |
+|-----------|----------|
+| Skills | `SKILL.md` |
+| Memory | `CLAUDE.md`, `AGENTS.md` |
+| Hooks | `${STATE_DIR}/settings.json` (Claude: .opencode/, OpenCode: .opencode/, Codex: .codex/) |
+| MCP | `*.mcp.json` |
+| Cursor | `.cursor/rules/*.mdc` |
+| Copilot | `.github/copilot-instructions.md` |
+
+## Error Handling
+
+- **agnix not installed**: Show install command `cargo install agnix-cli`
+- **Invalid path**: Exit with "Path not found: [path]"
+- **Parse errors**: Show raw agnix output
+
+## Links
+
+- [agnix GitHub](https://github.com/avifenesh/agnix)
+- [Rules Reference](https://github.com/avifenesh/agnix/blob/main/knowledge-base/VALIDATION-RULES.md)

package/adapters/opencode/commands/audit-project-agents.md

@@ -0,0 +1,125 @@
+<!-- AUTO-GENERATED by scripts/gen-adapters.js - DO NOT EDIT -->
+---
+description: "Use when coordinating multi-agent review passes in /audit-project. Details agent specialization, file filtering, and review queue handling."
+agent: general
+---
+
+> **OpenCode Note**: Invoke agents using `@agent-name` syntax.
+> Available agents: task-discoverer, exploration-agent, planning-agent,
+> implementation-agent, deslop-agent, delivery-validator, sync-docs-agent, consult-agent
+> Example: `@exploration-agent analyze the codebase`
+
+
+# Phase 2: Multi-Agent Review - Reference
+
+This file contains detailed agent coordination for `/audit-project`.
+
+**Parent document**: `audit-project.md`
+
+**Review Pass Definitions**: See `orchestrate-review` skill for canonical pass definitions (core + conditional). This command uses the same review passes but detects signals from project structure (not just changed files).
+
+## Agent Specialization
+
+### File Filtering by Agent
+
+Each agent reviews only relevant files:
+
+| Agent | File Patterns |
+|-------|--------------|
+| code-quality-reviewer | All source files (includes error handling) |
+| security-expert | Auth, validation, API endpoints, config |
+| performance-engineer | Hot paths, algorithms, loops, queries |
+| test-quality-guardian | Test files + missing-test signals |
+| architecture-reviewer | Cross-module boundaries, core packages |
+| database-specialist | Models, queries, migrations |
+| api-designer | API routes, controllers, handlers |
+| frontend-specialist | Components, state management |
+| backend-specialist | Services, domain logic, queues |
+| devops-reviewer | CI/CD configs, Dockerfiles |
+
+## Review Queue File
+
+Create a temporary review queue file in the platform state dir. Review passes append JSONL or return JSON for the parent to write.
+
+*(JavaScript reference - not executable in OpenCode)*
+
+## Agent Coordination
+
+Use Task tool to launch agents in parallel:
+
+- Invoke `@review` agent
+- Invoke `@review` agent
+- Invoke `@review` agent
+- Invoke `@review` agent
+- Invoke `@review` agent
+- Invoke `@review` agent
+- Invoke `@review` agent
+- Invoke `@review` agent
+- Invoke `@review` agent
+- Invoke `@review` agent
+
+
+## Finding Consolidation
+
+After all agents complete:
+
+*(JavaScript reference - not executable in OpenCode)*
+
+## Queue Cleanup
+
+After fixes and re-review, remove the queue file if no open issues remain:
+
+*(JavaScript reference - not executable in OpenCode)*
+
+## Framework-Specific Patterns
+
+### React Patterns
+
+*(JavaScript reference - not executable in OpenCode)*
+
+### Express Patterns
+
+*(JavaScript reference - not executable in OpenCode)*
+
+### Django Patterns
+
+*(JavaScript reference - not executable in OpenCode)*
+
+## Pattern Application
+
+*(JavaScript reference - not executable in OpenCode)*
+
+## Review Output Format
+
+```markdown
+## Agent Reports
+
+### security-expert
+**Files Reviewed**: X
+**Issues Found**: Y (Z critical, A high)
+
+Findings:
+1. [Finding details with file:line]
+2. [Finding details with file:line]
+
+### performance-engineer
+**Files Reviewed**: X
+**Issues Found**: Y
+
+Findings:
+1. [Finding details with file:line]
+
+[... per agent]
+
+## Consolidated Summary
+
+**Total Issues**: X
+- Critical: Y (must fix)
+- High: Z (should fix)
+- Medium: A (consider)
+- Low: B (nice to have)
+
+**Top Files by Issue Count**:
+1. src/api/users.ts: 5 issues
+2. src/auth/session.ts: 3 issues
+```