glm5-parallellisted

# GLM-5 Parallel Skill Model-agnostic parallel execution with Agent Teams coordination for comprehensive task execution. > **Note**: Despite the name, this skill is model-agnostic as of v2.88.0. It works with any model configured in Agent Teams. ## Usage ``` /glm5-parallel <task> [--roles coder,reviewer,tester] ``` ## Example ``` /glm5-parallel "Implement OAuth2 authentication" --roles coder,reviewer,tester ``` This spawns 3 teammates: 1. **coder** - Implements OAuth2 2. **reviewer** - Reviews the implementation 3. **tester** - Generates tests ## Execution Pattern ```bash # Generate task ID TASK_ID="parallel-$(date +%s)" # Parse roles ROLES="coder,reviewer,tester" # Spawn in parallel for ROLE in $(echo $ROLES | tr ',' ' '); do .claude/scripts/glm5-teammate.sh "$ROLE" "<task>" "${TASK_ID}-${ROLE}" & done # Wait for all wait # Aggregate results for ROLE in $(echo $ROLES | tr ',' ' '); do echo "=== $ROLE ===" cat .ralph/teammates/${TASK_ID}-${ROLE}/status.json | jq '.output_summary' done ``` ## Default Roles If no `--roles` specified: - `coder` - Implementation - `reviewer` - Code review - `tester` - Test generation ## Output Creates parallel status files: - `.ralph/teammates/{task_id}-coder/status.json` - `.ralph/teammates/{task_id}-reviewer/status.json` - `.ralph/teammates/{task_id}-tester/status.json` Aggregated in `.ralph/team-status.json` ## Agent Teams Integration (v2.88) **Optimal Scenario**: Pure Agent Teams (Native) This skill uses Pure