debug-agent

A command for finding and fixing problems in systems where several AI agents work together.

In plain words
What is it for?
Use it to investigate agent loops, wrong routing, lost state, tool failures, timeouts, and inconsistent results.
Why use it?
It provides a structured way to trace unexpected behavior, identify root causes, and verify fixes.

Command

Part of the agent-architect plugin — 13 skills, 1 command, 5 MCP servers shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/latestaiagents/agent-skills/debug-agent
Clone the repo
git clone --depth 1 https://github.com/latestaiagents/agent-skills

Or install agent-architect, the plugin that ships this one along with the rest of its 13 skills, 1 command, 5 MCP servers.

Per session 8 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 348 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00008 $0.00348
Opus 5 $0.00004 $0.00174
Sonnet 5 $0.00002 $0.00070
Haiku 4.5 $0.00001 $0.00035

Measured 2d ago against content hash 7f90d74322bf, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

debug-agent scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/agent-architect/commands/debug-agent.md · 68 lines

What it actually says

/debug-agent

Diagnose and fix issues in your multi-agent system.

What I Need

Tell me:

  • What's the unexpected behavior?
  • Which agent(s) are involved?
  • Do you have traces/logs?
  • Is it reproducible?

Debug Process

Step 1: Identify Failure Point

I'll help locate:

  • Which agent failed
  • At what step in the workflow
  • What was the input/output

Step 2: Analyze Root Cause

Common issues:

  • Prompt ambiguity
  • Tool execution failures
  • State corruption
  • Infinite loops
  • Token limit exceeded

Step 3: Trace Analysis

If using LangSmith:

from langsmith import Client
client = Client()
runs = client.list_runs(project_name="your-project", error=True)

Step 4: Fix & Verify

I'll provide:

  • Specific fix for the issue
  • Test case to verify
  • Prevention measures

Common Issues

Symptom Likely Cause Fix
Agent loops forever No exit condition Add max_iterations
Wrong agent selected Poor routing Improve router prompt
State lost Missing checkpointer Add persistence
Timeout Agent stuck Add timeout + fallback
Inconsistent results Non-deterministic Set temperature=0

Debug Checklist

[ ] Check agent prompts for ambiguity
[ ] Verify tool definitions are correct
[ ] Inspect state at each transition
[ ] Review routing logic
[ ] Check for token limits
[ ] Verify checkpointing works
[ ] Test edge cases
Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 68 lines · 8 tokens per session scan A 7f90d74322bf

Subscribe to this mod's changes

debug-agent is a command published in the GitHub repository latestaiagents/agent-skills (5 stars, last pushed 4mo ago), licensed MIT. It adds 8 tokens to every session and 348 once invoked, about $0.0000 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other commands, from other repositories

fw-setup-use

Workspace FDK stack switch — nvm use from .nvmrc or explicit 10/9 (Node 24.11 vs 18); optional --write-nvmrc and --global flags.

freshworks-developers/fw-dev-tools · 45 tokens

fdk-migrate

Migrate a legacy Freshworks app from platform-version 2.3 (2.x), FDK 9.x, and Node 18 to Platform 3.0 with FDK 10.0.1 and Node 24.11.0. Transforms manifest, request templates, Crayons UI, and validates on the new toolchain. Does not install fdk/Node—use fw-setup (or prompt user to add it) per SKILL.md.

freshworks-developers/fw-dev-tools · 102 tokens

fdk-fix

Fix all platform validation errors and lint errors in a Freshworks app. Runs fdk validate, identifies fatal errors and lint issues, applies fixes iteratively, and re-validates until all errors are resolved.

freshworks-developers/fw-dev-tools · 46 tokens

fdk-react-create

Create a new Platform 3.0 React Meta app (default UI stack). Uses fdk create react-starter-template or react-meta skeletons with DEW components, metaConfig in manifest.json, and React Router.

freshworks-developers/fw-dev-tools · 49 tokens

fw-setup-status

Check FDK and Node.js installation status (/fw-setup status); optional --verbose diagnostics.

freshworks-developers/fw-dev-tools · 23 tokens

fdk-refactor

Reduce function complexity in a Freshworks app to meet cyclomatic complexity ≤ 7 per function. Extracts helper functions, simplifies conditionals, and preserves behavior while improving code quality.

freshworks-developers/fw-dev-tools · 41 tokens