Paradigm

Build a new external service integration as a packaged Python CLI. The skill guides the project layout, API client, command-line interface, credential declarations, health check, and tests so the tool can be installed and used by agents.

• Create a tool for a new external API
• Build a Python client for a service
• Add a CLI command for an integration

Retrieve standalone DocSend documents or files in a DocSend Space for inspection, summarization, or download. The workflow handles email verification, passcodes, resumable sessions, and folder traversal, while reporting files that could not be fetched.

• Download a standalone DocSend document
• Retrieve files from a DocSend Space
• Resume a document request after email verification

Check each discovered live tool CLI by running its standard `health` command through the sandbox credential path. Review a compact Slack friendly report showing pass, failure, or partial results, with brief evidence for each tool.

• Checking authentication and connectivity for live tools
• Validating brokered credential injection
• Preparing a Slack ready tool health report

Verify that a running deployment works through the user facing agent path. The checks cover tool loading, Slack messages and file transfers, company context, logs, and metrics, then produce a compact pass, fail, or partial report with evidence.

• Smoke testing a running deployment from Slack
• Checking deployment readiness before promotion
• Verifying Slack, files, context, logs, and metrics end to end

Develop and debug Centaur harness-server backends: add harness CLIs, normalize their output into the shared protocol, and investigate streaming, interrupts, session resume, model selection, and reasoning controls across supported harnesses.

• Add support for a new harness CLI
• Debug normalized output from Claude Code or Codex
• Investigate streaming, interruption, or session resume behavior

Use an authenticated Centaur MCP deployment to find and call team-approved tools. It checks identity when permissions matter, follows the deployment’s current CLI help, and summarizes consequential changes with relevant identifiers or links.

• Finding the right approved Centaur tool
• Running an unfamiliar Centaur CLI command
• Checking the signed-in user's permissions

Investigate Centaur incidents using VictoriaLogs for runtime logs and VictoriaMetrics for deployment and health metrics. The skill correlates a narrow time window with relevant code and workflow status, then reports a supported cause or bounded hypotheses and the evidence still needed.

• Investigating a stalled workflow or missing post
• Checking thread and execution evidence
• Reviewing tool-call or sandbox activity

Use this skill to translate a Console administrator’s request into Centaur permission changes. It can create supported principals and secrets, grant or revoke access, adjust Slack and sandbox permissions, and manage Slack Connect allowlists while checking for unsafe personal credential sharing.

• Grant a Slack channel access to a shared tool credential
• Create a principal for a Slack channel, user, or workflow
• Adjust a channel’s access to another Slack channel

Preview edits to a Centaur persona, overlay, or base system prompt without deploying them. The skill recomposes the prompt in a scratch directory and runs a fresh one-shot child agent, so you can compare its replies to the intended behavior.

• Try a revised persona prompt before deployment
• Compare agent replies before and after a prompt edit
• Check a base or overlay prompt with a fresh child agent

Create, run, and debug Centaur end-to-end tests under e2e/. Exercise the Slack path across services with scripted model responses, inspect observable results, and trace failures through stack logs, session records, and captured model requests.

• Add a scenario that checks behavior across Centaur services
• Run the local end-to-end stack and test suite
• Debug a failing end-to-end CI job

Investigate authentication, permission, credential, OAuth, token, secret, and API proxy failures by checking live VictoriaLogs evidence first. Summarize relevant status codes and error details, and distinguish upstream denials from local proxy or secret resolution problems.

• Investigate a 401, 403, or 503 response
• Trace an integration authentication failure
• Check why a credential or secret is not resolving

Answer internal history and project context questions by searching indexed company information and checking relevant live sources such as Slack, Linear, Google Docs, Drive, Calendar, or Granola. Reconcile the evidence and cite dates or source links when available.

• Find what was discussed or decided internally
• Check a project's prior context or current status
• Search meeting notes and calendar events

Prepare repository changes for review by creating a named branch, committing with a Conventional Commit message, pushing it, and opening a GitHub pull request. The workflow also covers adding required attribution and returning the PR link.

• Opening a pull request for repository changes
• Pushing an authorized branch to GitHub
• Updating an existing pull request