Manage rememora persistent memory — save knowledge, search prior context, or check status.
ALWAYS use this skill BEFORE any of these actions: (1) Implementing something non-trivial — search for prior decisions and patterns in this area before writing code. (2) Encountering unfamiliar code or architecture — search for entity knowledge before exploring from scratch. (3) The user references past work, past decisions, or "what we decided" — search for that context. (4) You are stuck or blocked on a problem — search for related cases and solutions. (5) Making a decision that could conflict with a prior one — search to check for consistency. This skill fires BEFORE you act — search first, then implement.
Use this skill when working on Rememora development tasks such as picking approved tickets, planning implementations, writing code in isolated worktrees, running validation, and preparing pull requests. This is the Codex-facing equivalent of the Claude Code rememora-developer subagent.
Use this skill when triaging the Rememora project board, assessing issue readiness, checking dependencies, categorizing work, and proposing or applying board status updates. This is the Codex-facing equivalent of the Claude Code rememora-triage subagent.
ALWAYS use this skill after ANY of these events during your work: (1) You or the user made an architectural or design decision — choosing a database, framework, algorithm, API design, or trade-off. (2) You fixed a non-trivial bug — especially silent failures, race conditions, configuration gotchas, or environment-specific issues. (3) You discovered a codebase pattern — a naming convention, error handling approach, testing strategy, or architectural idiom. (4) You discovered an important entity — a service, API endpoint, config file, environment variable, or integration point. (5) The user corrected your approach or stated a preference. This skill fires AUTOMATICALLY — do not wait for the user to ask. Save knowledge AS you work.