agentleFS
ScottRBKScottRBK/forgetful/skills/forgetful-cli-setup/SKILL.md

forgetful-cli-setup

Skill341 starsChanged 23 days agoMITView on GitHub

Set up the Forgetful CLI and connect from a terminal — install, local or remote mode, auth, and verification. Use when connecting a human or headless agent via shell, wiring CI with token auth, or operating a local server (serve, database selection, feature flags, re-embedding). Also covers the machine contract: --json output, exit codes, and the discovery ladder.

  • Reads credentials
  • Installs packages

The file

We split this skill 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:

    After changing embedding provider or model: forgetful re-embed --dry-run to preview, then forgetful re-embed (batch size tunable via …

    line 80

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill 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
SKILL.md

Setting up the Forgetful CLI

The CLI is the shell-native surface: same tool registry as MCP, reachable by anything that can run a command. Setup ends with a verified round-trip, not an installed binary.

Install

uvx forgetful-ai --version     # ephemeral, no install
pip install forgetful-ai       # or: uv tool install forgetful-ai

Both forgetful and forgetful-ai invoke the same CLI.

Choose a mode

Connection precedence: --local > --server URL > FORGETFUL_SERVER env > local default.

  • Local (default): runs in-process against the locally configured database — zero config, no server needed. Right for a personal knowledge base on one machine.
  • Remote: drives a deployed server. Interactive use: forgetful auth login --server <url> (browser OAuth; tokens cached under ~/.config/forgetful/, server written to the user config). Headless/CI use: set FORGETFUL_SERVER and FORGETFUL_TOKEN (bearer) in the environment — no browser involved.

Config merges shell env over ~/.config/forgetful/.env. Check state anytime with forgetful auth status.

Verify — the completion criterion

forgetful tools list                          # registry categories come back
forgetful project list --json                 # one real call round-trips

Done when: both succeed against the intended target (local or the remote URL).

The machine contract

For scripts and agents, pass --json on data commands — tools, call, memory, and project verbs: results are machine JSON, errors arrive as {"error": ...} on stderr. auth subcommands are human-oriented and don't take --json. Exit codes are contractual — 0 success, 1 tool or runtime error, 2 usage error. Parse output, branch on exit code.

The discovery ladder

  1. Curated verbs for the common paths: forgetful memory search|save|get|recent, forgetful project list.
  2. forgetful tools list [--category <c>] to discover every registry operation.
  3. forgetful tools info <operation> for the schema.
  4. forgetful call <operation> --args '<JSON>' to invoke anything the registry has.

The judgment for what to store, query, and link lives in the domain skills (forgetful-remember, forgetful-recall, ...) — the ladder is how the CLI reaches the same operations.

Operating a local server

  • forgetful serve --transport http --host 0.0.0.0 --port 8020 (or --transport stdio for MCP clients that spawn a process).
  • Database via env: DATABASE=SQLite with SQLITE_PATH (file) or SQLITE_MEMORY=true (ephemeral), or DATABASE=Postgres with POSTGRES_HOST/PORT/DB/USER/PASSWORD.
  • Feature flags (default off): SKILLS_ENABLED, FILES_ENABLED, PLANNING_ENABLED — discovery reflects whatever is enabled.
  • After changing embedding provider or model: forgetful re-embed --dry-run to preview, then forgetful re-embed (batch size tunable via --batch-size).

Done when: /health answers (HTTP) or an MCP client lists the three meta-tools (stdio).

Planning write warnings

For generic call create_plan, call update_plan and call transition_task, inspect nonempty plugin_warnings while retaining the flat success payload. plan_auto_completion_rejected and plan_auto_completion_unavailable mean the task succeeded but automatic parent completion did not. Read current task and plan state before deciding on another action; do not blindly retry a committed task. Direct administrator policy errors use PLAN_PLUGIN_* codes. Local and remote modes expose the same contract. No new plan verbs or deletion MCP tools exist; plan/task deletion stays REST-only. Do not send automatic as an argument.

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.