Where most memory tools help your agent remember more, mnemoa helps it forget what no longer belongs

Description and use

mnemoa trims stale and redundant instruction files for coding agents, promoting rules to automated checks and keeping repository memory small and accurate.

Problem

Coding agents accumulate stale, redundant, and contradictory instructions in files like AGENTS.md and CLAUDE.md, leading to bloated context and drift from the codebase.

Outcome

mnemoa reduces context bloat by pruning stale instructions and promotes redundant rules to automated checks, keeping agent guidance concise and up-to-date.

Workflow

• Scan repository with mnemoa to identify issues
• Review the prune plan and token cost estimates
• Apply the prune plan using mnemoa forget --apply --yes
• Optionally promote redundant rules to automated checks and record in ledger
• Integrate mnemoa scans into CI via GitHub Action

Evidence

mnemoa scans a repository’s agent instruction files and plans. It checks for issues such as: Instructions that are too long or consume too much always-loaded context, Broken paths, symbols, and commands, Plans that have drifted from the code, Rules that a formatter, linter, or other tool could enforce instead

Source

It can also create a prune plan showing what it would remove or move, along with the estimated token cost before and after. You can review the diff before applying it.

Source

mnemoa is written in Rust and runs deterministic checks locally. The optional LLM feature can help draft rules, but it is disabled by default.

Source