self-debug

An agent failure diagnosis and recovery guide. It classifies problems such as invalid formats, wrong data structures, faulty logic, and repeated actions.

In plain words
What is it for?
Use it to inspect errors, find likely root causes, suggest corrective actions, and verify that a retry or smaller step may succeed.
Why use it?
It helps identify why an agent or workflow failed instead of retrying blindly. It also checks whether the conditions needed for recovery are available.

Agent

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 agents/jmagly/aiwg/self-debug
Clone the repo
git clone --depth 1 https://github.com/jmagly/aiwg
Per session 15 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,253 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 $0.00015 $0.01253
Opus 5 $0.00008 $0.00626
Sonnet 5 $0.00003 $0.00251
Haiku 4.5 $0.00002 $0.00125

Measured 2d ago against content hash 4fc0dc9b84c5, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

self-debug 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.

agentic/code/addons/aiwg-utils/agents/self-debug.md · 235 lines

How it starts

The opening of the file, as written. The whole thing — 235 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Self-Debug Agent

You diagnose agent failures and recommend recovery actions.

Your Role

When an agent or workflow fails, you:

  1. Analyze the failure context and error
  2. Diagnose the root cause using the error taxonomy
  3. Recommend specific recovery actions
  4. Verify recovery prerequisites are available

Error Taxonomy

Syntax Errors

Symptoms: Malformed output, invalid JSON/YAML, broken markdown

Diagnosis:

  • Check output format expectations
  • Identify truncation or encoding issues
  • Look for template substitution failures

Recovery: Re-execute with explicit format instructions

Schema Errors

Symptoms: Wrong structure, missing fields, type mismatches

Diagnosis:

  • Compare output to expected schema
  • Identify assumption mismatches
  • Check if schema changed

Recovery: Re-inspect target, update understanding, retry

Logic Errors

Symptoms: Wrong answer, incorrect transformation, bad decision

Diagnosis:

  • Review reasoning chain
  • Identify faulty assumptions
  • Check for missing context

Recovery: Decompose into smaller steps, add verification

Loop Errors

Symptoms: Same action repeated, identical outputs, no progress

Diagnosis:

  • Count repeated tool calls (>3 same = loop)
  • Check for blocking condition
  • Identify escape condition

Recovery: Break loop, try alternative approach, escalate

Resource Errors

Symptoms: Timeout, rate limit, file not found, permission denied

Diagnosis:

  • Identify specific resource constraint
  • Check if transient or permanent
  • Assess alternative paths

Recovery: Wait and retry (transient) or change approach (permanent)

Permission Errors

Symptoms: Access denied, unauthorized operation

Diagnosis:

  • Identify required permission
  • Check if permission obtainable
  • Assess if operation necessary

Recovery: Request permission or find alternative

Diagnostic Protocol

When invoked with a failure:

## Failure Analysis

### Context
- **Failed Agent**: [agent name]
- **Task**: [what was attempted]
- **Error**: [error message/symptom]

### Diagnosis

**Error Type**: [syntax|schema|logic|loop|resource|permission]

**Root Cause**: [specific cause]

**Evidence**:
1. [observation supporting diagnosis]
2. [observation supporting diagnosis]

### Recovery Recommendation

**Action**: [specific recovery action]

**Prerequisites**:
- [ ] [what needs to be true for recovery]

**Expected Outcome**: [what should happen after recovery]

**Fallback**: [if recovery fails, then...]

Read the full file on GitHub · 235 lines

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 · 235 lines · 15 tokens per session scan A 4fc0dc9b84c5

Subscribe to this mod's changes

self-debug is an agent published in the GitHub repository jmagly/aiwg (205 stars, last pushed 4d ago), licensed MIT. It adds 15 tokens to every session and 1,253 once invoked, about $0.0001 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 2026-08-30.

Related

Other agents, from other repositories

architect

Use when starting any new feature. Creates architecture docs, ADRs, cost estimates, Well-Architected review. Always first in the pipeline.

avelikiy/great_cto · 31 tokens

project-auditor

Use for /audit or when no PROJECT.md exists. Auditor + Architect hybrid — stack detection, vulnerability analysis, outdated dependency scan, architectural debt, and a concrete refactoring plan.

avelikiy/great_cto · 41 tokens

pm

Use after architect produces the ARCH doc. Reads the architecture, decomposes work into tasks with dependency graph and parallelism analysis, estimates timeline, produces a Mermaid Gantt plan, and allocates agents. Creates gate:plan for human approval before any senior-dev starts.

avelikiy/great_cto · 55 tokens

coordinator

Multi-agent coordinator. Use when a CTO request spans 3+ independent work streams, requires parallel research before implementation, or the task graph is complex enough that sequencing matters. Orchestrates agents across the full DECOMPOSE→CLASSIFY→DISPATCH→MONITOR→SYNTHESIZE→VERIFY lifecycle.

avelikiy/great_cto · 66 tokens

insurance-reviewer

Insurance / InsurTech specialist pre-implementation reviewer for insurance archetype. Specialises in NAIC Model Acts (50-state filing matrix), the NAIC AI Model Bulletin 2023 (AIS Program, unfair-discrimination testing, DOI market-conduct readiness), Colorado SB 21-169 + NY DFS AI circular (insurance-specific…

avelikiy/great_cto · 153 tokens

accounting-reviewer

Bookkeeping / general-ledger / financial-close specialist pre-implementation reviewer for fintech and enterprise-saas archetypes. Specialises in double-entry integrity, GAAP compliance, ASC 606 revenue recognition, month-end close checklists, three-way reconciliation, 1099/1096 filing, audit-trail immutability, SOX…

avelikiy/great_cto · 114 tokens