Skip to main content

A reasonably sized autonomous AI construction kit

Project description

Docketeer

The core agent engine for building autonomous AI assistants with Docket.

Docketeer is a small, opinionated toolkit for running an AI agent that can manage its own memory, schedule its own future work, and extend itself through plugins. The inference backend is pluggable — bring your own LLM provider. The core package provides the agent loop, a persistent workspace for the agent's files, and a plugin system based on standard Python entry points.

Tools

Workspace

  • list_files — list files and directories in the workspace
  • read_file — read contents of a text file
  • write_file — write content to a text file
  • edit_file — search-and-replace editing within a file
  • delete_file — delete a file
  • create_link — create a symbolic link in the workspace
  • read_link — read the target of a symbolic link
  • search_files — semantic search across workspace files (falls back to keyword grep without docketeer-search)

Scheduling

  • schedule — schedule a future nudge to prompt the agent at a given time
  • schedule_every — schedule a recurring nudge on an interval
  • cancel_task — cancel a scheduled task
  • list_scheduled — list all scheduled and running tasks

Chat

  • list_rooms — list available chat rooms
  • room_messages — read recent messages from a room
  • send_message — send a message to a room
  • react — react to a message with an emoji
  • wrap_up_silently — end a turn without replying

Vault

  • list_secrets — list stored secret names
  • store_secret — store a secret by name
  • generate_secret — generate and store a random secret
  • delete_secret — delete a stored secret
  • capture_secret — capture a secret from command output

Executor

  • run — run a command in the sandbox
  • shell — run a shell command in the sandbox

Configuration

Variable Default Description
DOCKETEER_DATA_DIR ~/.docketeer Where the agent stores memory and audit logs
DOCKETEER_DOCKET_URL redis://localhost:6379/0 Redis connection for task scheduling
DOCKETEER_DOCKET_NAME docketeer Name of the Docket instance
DOCKETEER_CHAT_MODEL balanced Model tier for chat conversations
DOCKETEER_LOG_LEVEL INFO Logging level (DEBUG, INFO, WARNING, ERROR, CRITICAL)
DOCKETEER_CHAT (auto) Entry point name to select when multiple chat plugins are installed
DOCKETEER_INFERENCE (auto) Entry point name to select when multiple inference plugins are installed
DOCKETEER_EXECUTOR (auto) Entry point name to select when multiple executor plugins are installed
DOCKETEER_VAULT (auto) Entry point name to select when multiple vault plugins are installed
DOCKETEER_SEARCH (auto) Entry point name to select when multiple search plugins are installed

Plugins

Docketeer discovers plugins through these entry point groups:

  • docketeer.inference — inference backends (which LLM provider powers the agent)
  • docketeer.chat — chat backends (how the agent talks to people)
  • docketeer.executor — command executors (sandboxed process execution)
  • docketeer.vault — secret vaults (store and resolve secrets)
  • docketeer.search — search catalogs (semantic search over workspace and tools)
  • docketeer.tools — tool plugins (what the agent can do)
  • docketeer.prompt — system prompt providers (contribute blocks to the system prompt)
  • docketeer.tasks — background task plugins (periodic or scheduled work)

Available plugins:

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

docketeer-0.0.17.tar.gz (108.9 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

docketeer-0.0.17-py3-none-any.whl (66.3 kB view details)

Uploaded Python 3

File details

Details for the file docketeer-0.0.17.tar.gz.

File metadata

  • Download URL: docketeer-0.0.17.tar.gz
  • Upload date:
  • Size: 108.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for docketeer-0.0.17.tar.gz
Algorithm Hash digest
SHA256 07593588f43b3be4e06aa165d6185f73548086d3e83a2af765f00954d4e5a955
MD5 0d002a0cfaa350d4064d739fee0468ce
BLAKE2b-256 afc7c82a70d5484f2297731e159ac648ee57b8e6dec0ce44fb948cec54941f2c

See more details on using hashes here.

Provenance

The following attestation bundles were made for docketeer-0.0.17.tar.gz:

Publisher: publish.yml on chrisguidry/docketeer

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file docketeer-0.0.17-py3-none-any.whl.

File metadata

  • Download URL: docketeer-0.0.17-py3-none-any.whl
  • Upload date:
  • Size: 66.3 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for docketeer-0.0.17-py3-none-any.whl
Algorithm Hash digest
SHA256 31e9846112a5b25d5480f3229101b55a51d060bbddfa1eafcf7531cda09b26d5
MD5 f638407e6e5f55f0b74f27088b0c09e1
BLAKE2b-256 177ebc3caca8a03b35cb34ed3cca58fc763bd9381369c87429c9c48f43e39e1d

See more details on using hashes here.

Provenance

The following attestation bundles were made for docketeer-0.0.17-py3-none-any.whl:

Publisher: publish.yml on chrisguidry/docketeer

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page