resolve-run

Resolve agent sessions into structured JSON evidence manifests with transcripts and logs.

1|Updated Mar 24, 2026

Search Tags:#debugging #agent #evaluation #orchestration #observability #asana

Ready to run

npx skills add https://github.com/EdgeApp/edge-dev-agents --skill resolve-run

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: resolve-run
Source: https://github.com/EdgeApp/edge-dev-agents/tree/main/.cursor/skills/resolve-run
Command: npx skills add https://github.com/EdgeApp/edge-dev-agents --skill resolve-run

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires jq, curl, node, tmux, git, and includes scripts (resource) components.

What problem does it solve?

This skill solves the difficulty of manually aggregating scattered logs, transcripts, and infrastructure states to evaluate the performance and outcomes of autonomous agent sessions.

Core Features & Use Cases

Evidence Aggregation: Automatically compiles transcripts, PR links, Asana status, and resource release receipts into a single JSON manifest.
Friction Analysis: Provides deep insights into agent struggle, including hook blocks, tool errors, and build invocations.
Use Case: Use this skill to generate a standardized evidence report for a specific task GID to feed into downstream evaluation tools or to audit agent behavior over a specific date range.

Quick Start

Run the resolve-run script with the task GID to generate a compact JSON evidence manifest for the specified agent session.

Frequently Asked Questions about resolve-run

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

FAQPage Schema

How do I generate an evidence manifest for autonomous agent sessions?▼

To generate an evidence manifest, run the resolve-run script with a specific task GID or date range to compile scattered transcripts, infrastructure logs, and Asana task metadata into a single structured JSON file.

What is the best way to audit autonomous agent behavior over a specific date range?▼

Auditing agent behavior over a date range is done by running the resolve-run script with a time-based window to reconstruct the full execution context into a standardized JSON evidence report.

How do I analyze agent friction like hook blocks and tool errors from orchestrated runs?▼

Analyzing agent friction involves resolving orchestrated agent sessions into an evidence manifest that highlights hook blocks, tool errors, and build invocations encountered during the autonomous run.

Do I need tmux and Asana API credentials to compile agent execution context?▼

Yes, compiling the full execution context requires access to local watcher logs, active tmux sessions, and Asana API credentials to accurately reconstruct infrastructure states and task statuses.

Can I use jq and curl to extract transcripts and PR links for downstream evaluation?▼

Yes, the skill leverages dependencies like jq, curl, and node to automatically aggregate PR links, transcripts, and resource release receipts into a compact JSON format suitable for downstream evaluation tools.

What are the limitations of resolving agent runs without local watcher logs?▼

Without local watcher logs, the resolve-run skill cannot fully reconstruct the execution context, limiting its ability to aggregate transcripts and infrastructure states into a comprehensive evidence manifest.