Classify coding agent memories and generate evidence bundles for audit replay.
npx skills add https://github.com/dmae97/oh-my-kimi --skill agentmemory
Please help me install this Agent Skill. Skill: agentmemory Source: https://github.com/dmae97/oh-my-kimi/tree/main/templates/skills/agents/agentmemory Command: npx skills add https://github.com/dmae97/oh-my-kimi --skill agentmemory
agentmemory helps you persist, govern, and audit what an AI agent chooses to remember so future runs can use reliable context without leaking secrets or stale assumptions.
Use the agentmemory skill to design what your agent should store in OMK project memory and produce an evidence bundle for a recent coding run.
High-intent search queries and answers about installing and using this skill.
Persistent agent memory for coding workflows involves deciding which project facts and decisions to store, then replaying them safely in future runs. This requires inventorying memory surfaces, classifying memory stability, and enforcing no-secrets guardrails to ensure reliable context recall.
Session replay for AI agent audits works by bundling prompts, generated diffs, verification outputs, and logs into evidence-focused artifacts. This creates a reproducible audit trail showing exactly what happened during a coding run, including limitations and verification steps consulted.
Memory governance in AI agent workflows is the enforcement of strict guardrails to prevent writing credentials or private data into memory. It ensures stored memories do not override live sources or official docs, maintaining safety and reliability across multiple runs.
Yes, you can use agent memory governance for project context recall by classifying candidate memories as project facts, decisions, commands, or risks. This reduces repeated prompting and ensures the agent retrieves reliable, governed context for coding tasks.
Limitations of replaying agent memory include refresh constraints and the risk of stale assumptions if memory is not updated. The system returns refresh limitations and verification artifacts to indicate when replayed context may be outdated or require validation against live sources.
jimmi2051Classify session findings into project memory, notepad, or working notes.
doolinAutomate a compliant GitHub Actions CI/CD pipeline with SSDF checks, SBOM generation, and OSCAL artifacts.
KentoShimizuAutomate browser-level verification with traces, screenshots, and replay logs.
BedaBeda-GrowthLaunch remote Jules sessions from approved deep plans and capture session IDs.
