Reusable capability bundles for OpenRoutines.
Each top-level directory is one plugin containing a PLUGIN.md manifest and its grouped routines, skills, and optional knowledge-ledger stubs. Plugins are executable supply-chain input: review the declared authority, routine prompts, and skill files before installing or updating one.
agentmail-report-- reports the agent's intentions and progress to an email address as a teammate-style update, sent from the agent's own AgentMail inbox.discord-report-- reports the agent's intentions and progress to a Discord channel as a teammate-style update, via a channel webhook.github-docs-- watches documentation repositories for changes, keeps a running account of what changed and what it affects, and flags docs filed in the wrong folder or carrying the wrong tags -- the base of a knowledgebot.github-issues-- conservatively triages a GitHub repository's issues and prints a weekly issue-health digest.slack-report-- reports the agent's intentions and progress to a Slack channel as a teammate-style update, via a minimal Slack app's bot token.steady-- connects an agent to Steady like a teammate: a daily check-in filed from the knowledge feed, and prompt replies to comments addressed to the agent.telegram-report-- reports the agent's intentions and progress to a Telegram chat as a teammate-style update, and answers replies to the report, via a BotFather bot token.
openroutines plugin add steadyspacecorp/openroutines-plugins --path github-issuesInstallation shows the plugin's grant summary, vendors it under .openroutines/plugins/github-issues/ with exact source provenance, and forces its routines inactive. Follow the plugin's setup instructions, review the diff, run openroutines check, and activate only the routines you want.
Installed plugins can later be inspected and updated from inside the agent:
openroutines plugin list
openroutines plugin update github-issuesUpdates validate and summarize the new upstream version before making changes, then three-way merge it with local edits. Newly added routines remain inactive, and plugin updates never overwrite live knowledge.
See AGENTS.md for the plugin format, security boundaries, and validation workflow.