openakari

Use when a file or module is too large, responsibilities are tangled, or a cross-cutting redesign is needed

Use when literature notes exist and need citation verification, or before publishing any artifact that cites literature

Use at the end of a work session to embed learnings into conventions. Simplified for models with lower reasoning capacity.

Use at the end of a work session, or when accumulated findings need to be embedded into conventions, skills, or patterns

Use when the Slack bot needs to handle an operational action like launching an experiment, querying status, or processing approvals

Use when a plan, finding, or design needs adversarial review before committing to it

Use when designing a new experiment or research protocol and need to ensure methodological rigor

Use when implementing new features, fixing bugs, or modifying code in infra/

Use when experimental results are unexpected, confusing, or need interpretation

Use when the PI or a human provides feedback, corrections, or direction on agent work

Use when a manual fix or workaround keeps recurring and might need to be formalized into infrastructure

Use when akari needs to proactively scan for new GenAI developments — model releases, capability changes, relevant research — rather than waiting for human input

Use when a research topic needs literature review, or when papers need to be found and documented for a project

Use at the start of a work session to select a task. Simplified for models with lower reasoning capacity.

Use at the start of every work session to assess current state and select the highest-value task

Use when an agent produced a flawed output and you need to understand why the reasoning went wrong

Create new research projects — either propose (agent-initiated gap analysis → formal proposal) or scaffold (human-initiated interview → project directory)

Use when a paper draft exists and needs to be prepared for venue submission or arxiv preprint

Use when skills may be out of date with source code, after significant infra changes, or when skill descriptions need CSO compliance audit

Use when a status report, research digest, or experiment comparison is needed for human review

Validate experiment metrics and findings — run metrics-first to check computations are meaningful, then findings to check conclusions are valid

Use when sessions have committed code and you need to check compliance with CLAUDE.md conventions, log entry standards, and experiment validation

Use when a plan or implementation feels over-engineered and needs a complexity check

Use when a Slack thread shows the bot behaving incorrectly and the root cause needs investigation

Use when multiple experiments or analyses have accumulated and their findings need to be interpreted together