What information is included in the Execution Report?

The report includes subagent status, verified token totals, commit summaries, files changed, and any issues discovered during the run.

How does it handle context compactions?

It parses the session history to sum tokens across all events, ensuring that usage from before a compaction is included in the final report.

Why is token metric verification important for subagents?

Accurate verification ensures the AI reports actual measured usage rather than estimates, which is critical for cost tracking and understanding context pressure during complex tasks.

Can I use this skill if a subagent hasn't finished?

Yes, the skill is designed to handle partial collections if a subagent is stalled, hits context limits, or requires manual intervention.

Does this skill automatically merge code?

No, this skill collects and prepares results; the actual merging of branches is typically handled by the related cat:merge-subagent skill.

Subagent Result Collector

Name: Subagent Result Collector
Author: cowwoc

bycowwoc

•

Productivity & Workflow

Extracts and reports work products, commit history, and token metrics from completed or interrupted Claude Code subagents.

The Subagent Result Collector is a vital workflow utility that streamlines the integration of subagent tasks back into parent branches. It automates the extraction of commit history, file changes, and discovered issues while providing critical visibility into subagent performance. By parsing cumulative token usage across context compactions and generating detailed execution reports, it ensures that developers have accurate data to validate quality and efficiency before merging work.

Key Features

01Accurate cumulative token metric parsing that spans all context compaction events

02Automated extraction of git commit history and code diffs from subagent worktrees

03Lightweight verification via completion markers to minimize session overhead

04Automated parent state updates for seamless task tracking and workflow management

05Comprehensive execution reporting including status, quality, and discovered issues

062 GitHub stars

Use Cases

01Integrating completed subagent feature work back into the main task branch

02Recovering partial progress from subagents that hit context limits or require intervention

03Auditing AI execution quality and token efficiency for complex code generations

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add cowwoc/claude-code-cat collect-results

For use in Claude.ai and ChatGPT

Download Skill