brewcode:convention

Mode Detection

Arguments: $ARGUMENTS

P0: Mode + Stack Detection

Parse $ARGUMENTS for mode keyword. Default = full. For paths mode: split comma-separated paths after keyword.

Step 0.1: Detect Stack

EXECUTE using Bash tool:

bash "${CLAUDE_SKILL_DIR}/scripts/convention.sh" detect-stack && echo "---DETECT-OK---" || echo "---DETECT-FAILED---"

Output: JSON {"stacks":[...],"primary":"...","build_file":"...","modules":[...]}.

Step 0.2: Scan Project

EXECUTE using Bash tool:

bash "${CLAUDE_SKILL_DIR}/scripts/convention.sh" scan && echo "---SCAN-OK---" || echo "---SCAN-FAILED---"

Output: JSON with source_dirs, file_counts, modules, total_files.

If total_files > 1000: Warn user, suggest paths mode.

Step 0.3: Setup Convention Directory

EXECUTE using Bash tool:

bash "${CLAUDE_SKILL_DIR}/scripts/convention.sh" setup && echo "---SETUP-OK---" || echo "---SETUP-FAILED---"

STOP if FAILED -- cannot proceed without output directory.

Step 0.4: Validate (rules mode only)

EXECUTE using Bash tool:

bash "${CLAUDE_SKILL_DIR}/scripts/convention.sh" validate && echo "---VALID---" || echo "---INVALID---"

If rules mode + INVALID -- Exit: "Run /brewcode:convention conventions first."

P1: Load Layer Definitions

Read references/analysis-layers.md. Filter layers by detected stack from P0. For paths mode: further filter by specified paths -- match layer file patterns against provided paths. Build ACTIVE_LAYERS for P2.

P2: Parallel Layer Analysis (10 agents, ONE message)

Dynamic Agent Resolution

Before spawning agents, check for project team agents:

1. If .claude/teams/ exists — read team.md for agent roster with domains
2. If team has architecture/testing domain agents — prefer over plugin architect/tester
3. Priority: team agent > project agent > plugin agent > system agent
4. If agent refuses (Task Acceptance Protocol) — re-delegate to suggested colleague (max 2 retries)

Always fall back to plugin agents when no project agents match the task domain.

Spawn ALL agents in a SINGLE message. Skip agents for inactive layers (filtered in P1).

Per-agent prompt template:

Analyze {LAYERS} in the project.
Stack: {DETECTED_STACK}

Layer definitions:
{LAYER_CRITERIA_FROM_ANALYSIS_LAYERS_MD}

Use grepai_search FIRST for file discovery, then Glob/Grep for verification.
For paths mode, scope analysis to: {SCOPED_PATHS}

Output format:

## Etalon Candidates
| Class | Path | Why Etalon | Score (1-10) |

## Naming Conventions
| Pattern | Example | Frequency |

## Directory Rules
| Rule | Path Pattern |

## Patterns (code snippets, max 3, 5-15 lines each)
### Pattern Name
` ```code```
Why: explanation

## Anti-Patterns
| Class | Problem | Fix |

P3: Etalon Selection (1 architect agent)

After all P2 agents complete, spawn 1 architect agent with combined results.

Prompt:

Receive analysis from 10 layer-analysis agents. Select 1-2 etalons per layer based on:
- Highest score from candidates
- Most complete pattern coverage
- Best naming convention adherence
- Fewest anti-patterns

If same class appears as etalon for multiple layers, assign to most relevant layer.

Input: {ALL_10_AGENT_OUTPUTS}

Output:
## Final Etalon Summary
| Layer | Etalon Class | Path | Score | Role |

## Conflict Resolutions
| Class | Claimed By | Assigned To | Reason |

## Coverage Gaps
| Layer | Issue | Recommendation |

P4: Document Generation (3 developer agents, PARALLEL)

Read references/conventions-guide.md for templates. Spawn 3 developer agents in ONE message.

Per-agent prompt:

Generate {DOCUMENT_NAME} following the template from conventions-guide.
Target structure: {TEMPLATE_FROM_CONVENTIONS_GUIDE}
Etalon selection: {P3_ETALON_SUMMARY}
Layer analyses: {RELEVANT_P2_OUTPUTS}
Stack: {DETECTED_STACK}
Write to: .claude/convention/{filename}.md
Structure: organized by layer -- each with Etalon Classes table, Patterns (5-15 lines each, max 3/layer), Anti-Patterns table, Quick Reference table at end.
Target: ~{LINE_COUNT} lines.

P5: Text Optimization

IF text-optimizer agent is available (brewtools installed):

Spawn 3 text-optimizer agents in ONE message (medium mode):

Task(subagent_type="text-optimizer", prompt="FIRST: Read $BT_PLUGIN_ROOT/skills/text-optimize/references/rules-review.md. THEN optimize .claude/convention/reference-patterns.md using medium mode. Output report with metrics.")
Task(subagent_type="text-optimizer", prompt="FIRST: Read $BT_PLUGIN_ROOT/skills/text-optimize/references/rules-review.md. THEN optimize .claude/convention/testing-conventions.md using medium mode. Output report with metrics.")
Task(subagent_type="text-optimizer", prompt="FIRST: Read $BT_PLUGIN_ROOT/skills/text-optimize/references/rules-review.md. THEN optimize .claude/convention/project-architecture.md using medium mode. Output report with metrics.")

ELSE (brewtools not installed -- fallback):

Read ${CLAUDE_SKILL_DIR}/../convention/references/text-optimize-fallback.md for compact rules. Apply rules manually to all 3 generated documents:

• .claude/convention/reference-patterns.md
• .claude/convention/testing-conventions.md
• .claude/convention/project-architecture.md

P6: User Review

Present summary:

## Convention Documents Generated

| Document | Lines | Key Etalons (top 5) |
|----------|-------|---------------------|
| reference-patterns.md | {N} | {class1}, {class2}, ... |
| testing-conventions.md | {N} | {class1}, {class2}, ... |
| project-architecture.md | {N} | {file1}, {file2}, ... |

AskUserQuestion options:

• A: Approve all -- continue to rules extraction
• B: Revise -- provide feedback (max 2 iterations, re-run P5 after ed