agentfeedback
AGENTS.md1 starsChanged 43 days agoMITView on GitHub
Asked to install this? Read AGENT-INSTALL.md (client) or AGENT-INSTALL-STACK.md (server and client). Both describe release 4.0.0 and later. Go + SQLite HTTP service that stores write-once feedback from AI coding agents (frictions, review runs, events) plus three skills: agentfeedback (submit/read on the agentfeedback CLI, installed everywhere), agentfeedback-triage (process the queue on the agentfeedback CLI plus an optional Python helper; user-invoked only) and agentfeedback-docs (generated reference docs, schemas and OpenAPI document for integrators and operators; skill render docs). Harness-neutral: this is…
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,542 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository6 other files it gives its agents.
- llms.txtllms.txt
- deliver-ticketSkill
- kaneo-cliSkill
- agentfeedback-docsSkill
- agentfeedbackSkill
- agentfeedback-triageSkill
The file
We split this AGENTS.md into the rules it gives an agent
2 rules. About 0 are common, 0 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 51
Read docs/develop.md: the commands, the rules that are not visible in the code, and the verification gates live there and only there.
- Very specific: line 53
Every delivered change lands in a vMAJOR.MINOR.PATCH release, which a maintainer cuts when they decide (docs/releases.md).
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
AgentFeedback: working in this repo
Asked to install this? Read AGENT-INSTALL.md (client) or AGENT-INSTALL-STACK.md (server and client). Both describe release 4.0.0 and later.
Go + SQLite HTTP service that stores write-once feedback from AI coding agents
(frictions, review runs, events) plus three skills: agentfeedback
(submit/read on the agentfeedback CLI, installed everywhere), agentfeedback-triage (process
the queue on the agentfeedback CLI plus an optional Python helper;
user-invoked only) and agentfeedback-docs (generated reference docs,
schemas and OpenAPI document for integrators and operators;
skill render docs).
Harness-neutral: this is the only agent instructions file (no CLAUDE.md or
GEMINI.md); Claude Code loads it when no CLAUDE.md exists.
Design records live in a private location; if .kitchen/ is present in this
checkout, read .kitchen/AGENTS.md first.
Contributor skills are versioned under .agents/skills/kaneo-cli/ and
.agents/skills/deliver-ticket/ (including its project sidecar); relative
links under .claude/skills/ expose the same copies to Claude Code.
These are development workflows, not skills shipped by the product installer.
A fresh clone needs no private planning files. Install kaneo-cli separately
following its skill's pinned-version guidance, and configure your own authorized
profile; credentials and instance URLs are never bundled. Delivery also needs
the tools listed in docs/develop.md, plus jq and gh.
The multi-model review skill is optional: delivery supports harness reviewers.
This file is for agents changing this repo. To use AgentFeedback,
read the README: the binary, agentfeedback init and the
generated agentfeedback skill, or
docs/api.md for raw HTTP. main describes release 4.0.0
and later; the documents of an earlier release are at its tag.
Route by task
| Task | Go to |
| Change service or client code | docs/develop.md: layout, commands, invisible rules, verification |
| Change the API | docs/openapi.yaml, schemas/ and conformance/ are the contract, docs/api.md its narrative; six artifacts move in one commit (see develop.md) |
| Change a skill or its scripts | internal/skillgen/source/ or skills/<name>/scripts/, then the skill gates in docs/develop.md; docs/recipes/http-curl.md and docs/recipes/http-powershell.md are generated from the same source |
| Change the session readers | docs/sessions.md |
| Change the install playbooks or the agent index | AGENT-INSTALL.md, AGENT-INSTALL-STACK.md, llms.txt; the artifact rule in develop.md |
| Change what the client collects, reads or writes | docs/security.md states it; keep it true |
| Change the integrations list or a hub artifact | integrations.json, schemas/integrations.v1.json, integrations/, docs/integrations/clawhub.md, docs/integrations/hermes.md, docs/integrations/nanoclaw.md; regenerate the README table with just box go test . -run TestIntegrationsTable -update |
| Change docs | Keep the README route table and this one true (just docs-check); one doc per task, no duplicated facts |
| Run, deploy, back up, migrate | docs/operate.md |
| Release | docs/releases.md |
Before you change anything
Read docs/develop.md: the commands, the rules that are not
visible in the code, and the verification gates live there and only there.
Every delivered change lands in a vMAJOR.MINOR.PATCH release, which a
maintainer cuts when they decide (docs/releases.md).
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.