gepa-optimize SKILL.md skillB
gepa-optimize from indexedlabs/pydantic-ai-gepa is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.
mdr add indexedlabs/pydantic-ai-gepa/gepa-optimize@git:20261014.8de3547Pins git:20261014.8de3547 in mdr.lock. Pin the fingerprint instead.
- Audit grade
- B, 16 of 17 checks passed
- Kind
- SKILL.md skill
- Repository
- indexedlabs/pydantic-ai-gepa
- Path
src/pydantic_ai_gepa/skills/gepa_optimize/SKILL.md- Versions recorded
- 9
- Latest version
- git:20261014.8de3547, committed Oct 14, 2026
- First version recorded
- Sep 26, 2026
- Size
- 79,030 bytes
- Repository stars
- 36
- License
- MIT
Optimize a pydantic-ai agent's instructions, tool descriptions, output type, and signature inputs using the gepa CLI. Trigger when the user asks to "optimize this agent", "improve tool descriptions", iterate on prompts driven by eval failures, or otherwise improve a pydantic-ai agent against a dataset. Operates in the user's repo with full filesystem access — YOU are the reflection model; the gepa CLI handles minibatches, evaluation, and Pareto bookkeeping.
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# gepa-optimize You are the reflection model. The `gepa` CLI is a small toolkit that handles minibatches, trace/report persistence, candidate evaluation, comparison, and history bookkeeping; you read failure reports and traces, edit component slots or source code, and continue until the run completes. There is no `propose` or `reflect` verb on the CLI because that's the work you do — while `gepa run` is paused, or between manual `gepa eval` invocations — by editing files. ## Driving an already-started run `gepa --gepa-dir /absolute/workspace run drive --run-id RUN --reflectors reflectors.json` automates lane runs and single-checkout held-out runs with one short-lived reflector per step. Run it in the harness environment. It leases/reset lanes, dispatches the fixed training-reflection prompt, guards process exit, and calls `run select` or `harness serve --once`. **No separate harness service, selector, or reflector may run for a driven run: the driver owns scoring timing.** The existing held-out lane refusal remains until standalone lane repositories land. The JSON configuration contains only allowed commands, for example: ```json …
Read the whole file at its exact version.
How to install
mdr add indexedlabs/pydantic-ai-gepa/gepa-optimize@sha256:f4076567c56ec36bThe line at the top of this page pins the label git:20261014.8de3547, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
[](https://markdownregistry.com/a/art_y4a7j5s7l7zpri4u)
3 badge image fetches in 30 days. What the badge shows.
Versions
The registry has recorded 9 versions of this file, between Sep 26, 2026 and Oct 14, 2026, about one every 2 days.
| version | committed | commit | size | audit | |
|---|---|---|---|---|---|
| git:20261014.8de3547 latest | 2026-10-14 | 8de3547 | 79,030 B | B | view · diff |
| git:20261014.1006022 | 2026-10-14 | 1006022 | 75,491 B | B | view · diff |
| git:20261007.4639ab0 | 2026-10-07 | 4639ab0 | 74,938 B | B | view · diff |
| git:20260927.6b92fdf | 2026-09-27 | 6b92fdf | 74,032 B | B | view · diff |
| git:20260926.1948eb9 | 2026-09-26 | 1948eb9 | 69,905 B | B | view · diff |
| git:20260926.76bba35 | 2026-09-26 | 76bba35 | 69,715 B | B | view · diff |
| git:20260926.fd05afe | 2026-09-26 | fd05afe | 62,853 B | B | view · diff |
| git:20260926.529a805 | 2026-09-26 | 529a805 | 59,771 B | B | view · diff |
| git:20260926.a0fb5f7 | 2026-09-26 | a0fb5f7 | 59,442 B | B | view |
Audit of the latest version
- fail: No instruction to read or print local credentials (matched: read from a nomination, `gepa.toml` or `.env)
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (79030 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Starting a new one? Use a template for SKILL.md on mdpantry, which passes every one of these checks and says what to change in it.
Source
indexedlabs/pydantic-ai-gepa · 36 stars · license MIT · pushed 2026-10-14 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_y4a7j5s7l7zpri4u GET https://markdownregistry.com/api/v1/resolve?ref=indexedlabs/pydantic-ai-gepa/gepa-optimize GET https://markdownregistry.com/api/v1/blob/f4076567c56ec36b6f3f211232c3f8cf94efdc44196cbac95349e7efb4180a38
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.