transcript-distillation

A workflow for turning coding-agent conversation logs into rows for training or evaluating AI models.

In plain words
What is it for?
Use it to process Claude Code or Codex logs, extract assistant turns, label their quality, score sessions by outcomes such as passing tests, and create SFT, KTO, DPO, or GRPO datasets.
Why use it?
It removes the manual work of reading transcripts, removing secrets and filler, and judging which exchanges were useful.

Skill for Claude CodeCodex

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 skills/profsynapse/synaptic-tuner/transcript-distillation
Any agent
npx skills add ProfSynapse/Synaptic-Tuner --skill transcript-distillation
Clone the repo
git clone --depth 1 https://github.com/ProfSynapse/Synaptic-Tuner

Made for: Claude Code, Codex.

Per session 136 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,617 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 $0.00136 $0.01617
Opus 5 $0.00068 $0.00809
Sonnet 5 $0.00027 $0.00323
Haiku 4.5 $0.00014 $0.00162

Measured 2d ago against content hash a701fa3e6d6b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

transcript-distillation 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.

The scan reads SKILL.md. This mod also ships 7 executable files (scripts/adapters/__init__.py, scripts/adapters/base.py, scripts/adapters/claude_code.py, …), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

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.

.agents/skills/transcript-distillation/SKILL.md · 137 lines

How it starts

The opening of the file, as written. The whole thing — 137 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Transcript Distillation

Mine local agent transcripts into training data. You point it at a directory of conversation logs; it emits one row per assistant turn, scrubs secrets, labels each turn (accept / borderline / reject), scores each session by how it ended up (tests passed, clean build, commit, PR, approval), drops ceremony/filler/ duplicates, and stamps a quality tier (gold/silver/bronze) on every row.

Format-agnostic. The engine knows nothing about Claude or Codex — that lives in small pluggable adapters. Two ship built-in (claude_code, codex); add more in ~40 lines (see reference/writing-adapters.md).

Quickstart

SKILL=.skills/transcript-distillation

# 1. copy the template config and (optionally) edit scope/paths
cp $SKILL/configs/template.yaml my_distill.yaml

# 2. smoke test on a few files first — ALWAYS do this before a full run
python $SKILL/scripts/distill.py --config my_distill.yaml --smoke --limit 20 --show 3

# 3. full run
python $SKILL/scripts/distill.py --config my_distill.yaml

Out of the box the template grabs all local Claude Code (~/.claude/projects) and Codex (~/.codex/sessions) transcripts and writes to ./transcript_distill_out/rows.jsonl. Only dependency: pyyaml.

CLI

Flag Meaning
--config PATH distill config YAML (required)
--smoke write to a smoke/ subdir (keeps full runs separate)
--limit N cap files per source (0 = all) — use for smoke tests
--sources a,b restrict to named sources from config
--show N print N sample rows to eyeball quality
--max-context-tokens N override context_budget.max_context_tokens for this run

Pipeline (all config-driven)

adapter.parse(file)            # format-specific -> normalized events
  -> one row per assistant turn
  -> secret scrub at emit time (raw secrets never hit disk)
  -> context filled most-recent-first under a TOKEN budget
  -> deterministic label: accept / borderline(->judge) / reject
  -> quality funnel: drop ceremony tools, trivial filler, duplicates
  -> session-outcome tiering: gold / silver / bronze
  -> JSONL

Read the full file on GitHub · 137 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 · 137 lines · 136 tokens per session scan A a701fa3e6d6b

Subscribe to this mod's changes

transcript-distillation is a skill published in the GitHub repository ProfSynapse/Synaptic-Tuner (27 stars, last pushed 2d ago), licensed MIT. It adds 136 tokens to every session and 1,617 once invoked, about $0.0007 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 skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

brainstorming

You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.

obra/superpowers · 37 tokens

chat-pet-sprite-creation

Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.

microsoft/vscode · 53 tokens

cpu-profile-analysis

Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…

microsoft/vscode · 71 tokens

babysit-pr

Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…

openai/codex · 114 tokens

imagegen

Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…

openai/codex · 113 tokens