A regression-feedback workflow that turns each escaped defect into a permanent test case, recording the model version and evidence involved.

Use it to record escapes, add them to a regression corpus, and rerun the corpus against the current pipeline.

It prevents the same pipeline mistake from being missed again and shows whether later versions still catch known failures.

feedback — Skill for Claude Code by fledgeling-co/fledgeling-plugins

Install

Getting it into your agent

It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.

Clone the repo

git clone --depth 1 https://github.com/fledgeling-co/fledgeling-plugins

agentmods

npx agentmods add skills/fledgeling-co/fledgeling-plugins/feedback

Made for: Claude Code.

Or install warrant, the plugin that ships this one along with the rest of its 9 skills.

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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/fledgeling-co/fledgeling-plugins/feedback/github.svg)](https://agentmods.dev/skills/fledgeling-co/fledgeling-plugins/feedback)

Your own site

<a href="https://agentmods.dev/skills/fledgeling-co/fledgeling-plugins/feedback"><img src="https://agentmods.dev/badge/skills/fledgeling-co/fledgeling-plugins/feedback/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.

Your own site · 80×15

<a href="https://agentmods.dev/skills/fledgeling-co/fledgeling-plugins/feedback"><img src="https://agentmods.dev/badge/skills/fledgeling-co/fledgeling-plugins/feedback.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 100 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 1,219 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.

Model	Per session	Once invoked
Fable 5.1	$0.00100	$0.01219
Opus 5.5	$0.00040	$0.00488
Sonnet 5.5	$0.00020	$0.00244
Haiku 4.5	$0.00010	$0.00122

Measured 1mo ago against content hash 54fb48dce57d, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

feedback 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 1mo 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/warrant/skills/feedback/SKILL.md · 111 lines

How it starts

The opening of the file, as written. The whole thing — 111 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Feedback — calibration from escapes

An earlier design ran a prospective reader study to establish a baseline. It was cut because it inverted the point: it spent human review time in order to remove human review time. This skill learns from the cases where the pipeline was wrong and someone said so.

That is cheaper and it never goes stale. A study measures one model version on one date; the regression corpus re-measures every version against every escape ever found, and it grows.

Running as a Gemini model? Read gemini.md in this directory first, then follow this file with the overrides it names. It forbids computing the escape rate that escape_report.py refuses to print, makes the corpus case count travel beside regress_run.py's exit code so an empty corpus cannot read as a tier-2 pass, and takes the recorded model version from lanes.toml rather than from what the run knows about itself. Other models skip it.

Procedure

Record the escape.
```
python3 scripts/feedback_record.py --root <repo> --class <name> --item <id> \
    --missed '<what the pipeline failed to identify>' --evidence-digest <digest> \
    --expected-verdict fail
```
It captures the defect class, the item, the warrant version, the model ids and versions that were live at the time, and the evidence digest the wrong verdict was written from. The digest is what makes the escape reproducible rather than anecdotal.
Build the regression case.
```
python3 scripts/regress_build.py --root <repo> --escape-id <id> --evidence-dir <dir>
python3 scripts/regress_build.py --root <repo> --all      # every escape not yet built
```
Writes .warrant/regression/<escape-id>/ with the inputs and the verdict that should have been returned. Idempotent, so re-running it on a known escape changes nothing.
Run the whole corpus again.
```
python3 scripts/regress_run.py --root <repo> --verdict-cmd '<command>'
```
Every historical escape runs against the current lanes. Exit 2 names each case no longer caught. This is the tier-2 entry condition: a class may be closed by machine only while the machine demonstrably catches everything it has previously missed in that class.

Read the full file on GitHub · 111 lines

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

gemini.md 15 KB

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.

1mo ago First seen · 111 lines · 100 tokens per session scan A 54fb48dce57d

Subscribe to this mod's changes

feedback is a skill published in the GitHub repository fledgeling-co/fledgeling-plugins (2 stars, last pushed 4d ago), licensed MIT. It adds 100 tokens to every session and 1,219 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-09-04.

E2E testing Testing

eval-dataset-design

A guide to creating reliable test sets for AI agents. Each evaluation task includes the request, the simulated user's behaviour, a reset starting state, and a success check that can be independently verified.

bojieli/ai-agent-book · 138 tokens

test-master

Generates test files, creates mocking strategies, analyzes code coverage, designs test architectures, and produces test plans and defect reports across functional, performance, and security testing disciplines. Use when writing unit tests, integration tests, or E2E tests; creating test strategies or automation…

Jeffallan/claude-skills · 104 tokens

verify-omnigent

Drive Omnigent the way a user does and prove a behavior with recorded evidence, using an isolated server, runner, and mock model plus a feature map of every user entry point. Load before reproducing a user-facing bug, before claiming a fix works, or when reviewing whether a change covered every surface (session vs.…

omnigent-ai/omnigent · 100 tokens

dogfood

Systematically explore and test a mobile app on iOS/Android with agent-device to find bugs, UX issues, and other problems. Use when asked to dogfood, QA, exploratory test, find issues, bug hunt, or test this app on mobile.

callstack/agent-device · 55 tokens

cua-driver

Use Cua Driver for desktop or browser tasks that are awkward or unavailable through Bash/APIs, or when the user explicitly wants GUI interaction: app testing, visual bug reproduction, form filling, calendar entry, screenshots, and demo recording. Also covers Cua setup; not OpenAI Codex Computer Use or web research.

davidondrej/skills · 68 tokens

critique

Adversarial design critique of the current work — render it, look at it, and argue for rejection. Run after the gates are green, never instead of them.

plugin87/ux-ui-agent-skills · 37 tokens