skills.rest
williamthorsen
williamthorsenCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

find-orchestration-savings

Analyze completed orchestrated runs for token waste and resource misallocation.

1|1|Updated Feb 25, 2026
Search Tags:#optimization#orchestration#subagent#token usage#cost analysis#run logs
One-click install
npx skills add https://github.com/williamthorsen/codeassembly --skill find-orchestration-savings-williamthorsen
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: find-orchestration-savings
Source: https://github.com/williamthorsen/codeassembly/tree/main/packages/agents/content/skills/find-orchestration-savings
Command: npx skills add https://github.com/williamthorsen/codeassembly --skill find-orchestration-savings-williamthorsen

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Orchestrated multi-agent runs can waste tokens and misallocate model resources, but without structured analysis of run logs these inefficiencies remain invisible and costs keep growing. ## Core Features & Use Cases - Run Log Analysis: Reads run-log.jsonl and run-index.json from a completed run directory to identify cost-saving opportunities. - Automated Context Resolution: Resolves the target run from session context or derives project, ticket, and artifact paths via the derive-session-context script when no path is given. - Subagent-Based Reporting: Dispatches a savings-analyzer subagent on the Haiku model that writes a findings artifact, then surfaces the top 3 findings in conversation. - Use Case: After finishing a large orchestrated code-generation run, invoke this Skill to discover which phases over-consumed tokens and where a cheaper model would have preserved output quality. ## Quick Start Analyze the most recent completed orchestrated run in this project and show me the top token-saving opportunities.

Frequently Asked Questions about find-orchestration-savings

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I analyze an orchestrated run for token waste?▼

Invoke the skill with the run directory path, or omit it to auto-resolve the most recent completed run from session context. The skill verifies run-log.jsonl and run-index.json exist, then dispatches a savings-analyzer subagent that writes a findings artifact.

What files are required to analyze an orchestration run?▼

The run directory must contain run-log.jsonl and run-index.json. If either file is missing, the skill reports the issue and exits without performing analysis.

Which model does the savings analysis use?▼

The savings-analyzer subagent runs on the Haiku model, configured via the savings_analyzer model key which defaults to haiku. This keeps analysis cost low relative to the runs being analyzed.

Can the savings analysis run automatically after an orchestrated run?▼

Yes, it auto-triggers during Phase 5 (summary) of orchestrated runs. The orchestrate engine dispatches the savings-analyzer as a background task while the orchestrator writes the run summary, adding zero latency.

What happens if no completed run is found for analysis?▼

The skill reports "No completed run found for this context" and exits. It scans the artifact directory for the current project and ticket, looking for the most recent completed run by timestamp.

Related Skills

Explore More Software Engineering Skills→
skurekjakubskurekjakub

agent-eval

Evaluate completed Ralph agent runs against scoring criteria and artifact contracts.

Community
Advanced
KamiJeongKamiJeong

speckit-token-observability

Track token usage across Spec Kit orchestration and workflow phases.

Community
Advanced
LiozShorLiozShor
2

find-session

Search local Claude Code session logs for session IDs, project directories, timestamps, and topic summaries.

Community
Advanced
Bonny94ITABonny94ITA

caveman-stats

Analyze Claude Code session logs to report token usage and savings.

Community
Basic