aesir

Plugin aesir from marketplace agentic-aesir in bpelotto/Agentic_Aesir. Commands read from the repo's manifest — 12 days ago.

Skills (2)

• playbook

  How the Aesir team works. Use when deciding which Aesir agent fits a task, planning a multi-step or multi-agent workflow, estimating what a task will cost in tokens, or adding and overriding team agents.

• setup

  Scaffold or refresh the Aesir setup. Project mode writes AGENTS.md, CLAUDE.md, and stack rules. --user writes baseline rules to ~/.claude/rules and syncs Copilot agents to ~/.copilot/agents. --copilot syncs Copilot files into this project.

Subagents (9)

• aegir

  Platform engineer. Use for infrastructure as code, containers, environments and secrets configuration, CI/CD pipelines, deployment gates, and release automation. Sizes pipelines to the project. Not for application code or code review.

• baldur

  Code and security reviewer. Use after code is written to review a diff, file, or PR for correctness, maintainability, tests, and security (access control, injection, secrets). Severity-ranked findings, no edits. Not for writing fixes.

• mimir

  Research and comparison. Use for technology or library evaluations, "how does X work" questions, news or trend digests, and dependency or CVE lookups that need web sources. Not for in-repo code search (use Explore) or for writing the final content.

• saga

  Writer and editor. Use for READMEs, API and developer docs, tutorials, blog posts, release notes, announcement and social copy, and editing passes on existing text. Edits files in place. Not for research (use mimir) or for code changes.

• thor

  Implementer. Use to build or change code in any layer (backend, frontend, data access, scripts, migrations) from a clear request or spec, including bug fixes. Builds and runs tests before reporting. Not for design decisions, reviews, or docs.

• tyr

  Product requirements writer. Use to produce PRDs, user stories with acceptance criteria, scope and non-goals, or to turn a vague idea into a testable spec in docs/specs. Not for task breakdown or sprint planning (use plan mode), design, or code.

• ull

  Tester. Use to plan coverage for a change or to write and run tests at any level: unit, integration, end-to-end, and basic load or benchmark checks. Reads existing tests first and reports real pass/fail results. Not for fixing the code under test.

• vidar

  Software architect. Use for system and API design, technology choices, architecture decision records, integration contracts, and risk or tradeoff analysis before building. Writes ADRs to docs/adr. Not for implementation code or routine refactors.

• volva

  AI and LLM specialist. Use for prompt and system-prompt design, agent and subagent architecture, choosing a model by capability and cost, evaluation design, and retrieval (RAG) pipelines. Not for general application code or non-AI research.

Similar plugins