analyze-conversation

analyze-conversation is a skill for Claude Code, Codex from genudo-ai/genudo_mcp. It costs 97 tokens per session (404 once invoked), scanned A, original, MIT.

A guide for investigating real Genudo conversations, where an AI agent chats with customers. It reconstructs the conversation and identifies where the agent went off course.

In plain words
What is it for?
Reviewing a conversation turn by turn, finding invented facts or incorrect actions, tracing the cause, and recording recommended fixes.
Why use it?
It helps separate what the agent did from the configuration that caused it, such as instructions, conversation stages, or actions.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Good fit Reviewing a conversation turn by turn, finding invented facts or incorrect actions, tracing the cause, and recording recommended fixes.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/genudo-ai/genudo_mcp/analyze-conversation
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.

Any agent
npx skills add genudo-ai/genudo_mcp --skill analyze-conversation
Clone the repo
git clone --depth 1 https://github.com/genudo-ai/genudo_mcp

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for analyze-conversation

README.md
[![agentmods](https://agentmods.dev/badge/skills/genudo-ai/genudo_mcp/analyze-conversation/github.svg)](https://agentmods.dev/skills/genudo-ai/genudo_mcp/analyze-conversation)
Your own site
<a href="https://agentmods.dev/skills/genudo-ai/genudo_mcp/analyze-conversation"><img src="https://agentmods.dev/badge/skills/genudo-ai/genudo_mcp/analyze-conversation/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for analyze-conversation

Your own site · 80×15
<a href="https://agentmods.dev/skills/genudo-ai/genudo_mcp/analyze-conversation"><img src="https://agentmods.dev/badge/skills/genudo-ai/genudo_mcp/analyze-conversation.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 97 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 404 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00097 $0.00404
Opus 5 $0.00048 $0.00202
Sonnet 5 $0.00019 $0.00081
Haiku 4.5 $0.00010 $0.00040

Measured 10d ago against content hash 227806b78c5d, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

analyze-conversation 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 10d 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.

chatgpt/genudo-chatgpt-plugin/skills/analyze-conversation/SKILL.md · 37 lines

What it actually says

Analyze Conversation

Reconstruct what an agent actually did in a real chat and explain why — then point to the fix.

Resolve the transcript

  • Given a conversation ID (from a link): fetch its messages with list_messages (filter by pipeline/contact as available; page with the cursor for long histories).
  • Given a phone number: list_contacts to resolve the contact, then list_messages filtered by contact_id.

Evaluate turn by turn

Walk the transcript in order and note, per turn:

  • which stage the agent was likely in,
  • whether it followed that stage's flow and honored global rules,
  • drift points — where it invented facts, claimed a false action success, asked for data it already had, skipped/entered a stage wrongly, or broke tone.

Root-cause

For each drift point, trace it to the responsible field: persona · global instructions · a stage's enter_condition · a stage's instructions · an action's instructions/variables. Cross-reference the live text with list_pipelines / list_pipeline_stages.

Output

Write ./genudo-work/conversations/<id>/analysis.md: a turn timeline, the drift points, the root cause per point, and the recommended fix. Hand off to diagnose-agent-behavioredit-pipeline-instructions. Read-only; no account writes here.

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. 10d ago First seen · 37 lines · 97 tokens per session scan A 227806b78c5d

Subscribe to this mod's changes

analyze-conversation is a skill published in the GitHub repository genudo-ai/genudo_mcp (0 stars, last pushed 10d ago), licensed MIT. It adds 97 tokens to every session and 404 once invoked, about $0.0005 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.