agentleFS
torisKRtorisKR/toris-agent/docs/AGENTS.md

toris-agent

AGENTS.md2 starsChanged 4 months agoApache-2.0View on GitHub

Toris has one agent catalogue. toris agents, TUI /agent, --agent, and the planner all read the same list. You do not fork the repo to add a domain specialist. Eleven task roles plus the toris chat persona live in src/core/agents.js. The planner can assign the task roles. toris is the session, not a plan step. One JSON object per file. Directories are not recursive (examples/ is ignored). Precedence matches skills: builtin < home < project. A project file with id…

The file

We split this AGENTS.md into the rules it gives an agent

1 rule. About 0 are common, 0 are distinctive, and 1 is 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.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Remove an overlay by deleting that JSON file.

    line 56

≈ “Common” is an estimate: at least 1 in 5 of the 42,283 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
AGENTS.md

Agent profiles — builtins and project-local roles

Toris has one agent catalogue. toris agents, TUI /agent, --agent, and the planner all read the same list. You do not fork the repo to add a domain specialist.

Builtins

Eleven task roles plus the toris chat persona live in src/core/agents.js. The planner can assign the task roles. toris is the session, not a plan step.

Overlay files

One JSON object per file. Directories are not recursive (examples/ is ignored).

~/.toris/agents/<id>.json          # optional user-global overlay
<repo>/.toris/agents/<id>.json     # project-local (wins on the same id)

Precedence matches skills: builtin < home < project. A project file with id implementer replaces the built-in Implementer. A new id appears everywhere the catalogue is shown and is assignable in plans (unless category is core). toris agents labels custom rows project or home, so you can see which file supplies the role.

There is no network fetch. Absent directories are empty, not errors. toris agents reports an invalid overlay with the file and field that need correction.

Schema

Required:

FieldRule
idSlug: ^[a-z][a-z-]*$. Must match the filename (aso-specialist.json).
titleNon-empty, max 80 characters.
categorycore | plan | build | review | verify | ship
writesJSON boolean. true if this role may edit files.
summary11–500 characters. Shown in pickers and used when system is omitted.

Optional:

FieldRule
systemSpecialist system prompt (max 8000 characters). Replaces the derived role prompt.

Unknown fields fail. Invalid JSON fails. The error names the file and the field — it does not crash with a stack trace. id toris may only use category: "core".

Manage and select a role

Copy aso-specialist.json to .toris/agents/aso-specialist.json to try a specialist locally. Create or edit overlays in your editor, then run toris agents to check the schema and inspect the effective catalogue. Keep the filename and id in agreement.

Remove an overlay by deleting that JSON file. If it replaced a home or builtin role, the lower-precedence role becomes available again on the next catalogue load. Use the SOURCE column to choose the file you intend to change.

toris agents
toris agents --json
toris --agent aso-specialist

In an interactive session, /agent lists the available roles and /agent aso-specialist selects one for subsequent turns. Start a new session after changing overlay files so it loads the updated catalogue.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.