Score breakdown
Popularity is tracked separately. Support, ads, sponsorships, and tips never affect these signals.
Why it matters
Useful for agent builders tired of re-teaching every session what the project calls things, how success is measured, and who owns what.
Where this stands now
Agent Memory Repo — Cognition's open spec for git-backed agent memory ranks #73 of 1507 tracked Agent Infrastructure items by composite score (7.1 against a section median of 4.5). The section currently carries 1419 Bronze, 57 Silver, 31 Gold. Signal extremes versus the section: momentum at the 94th percentile; novelty at the 95th percentile.
Who should use it
Who should skip it
Hold off on Agent Memory Repo — Cognition's open spec for git-backed agent memory until it graduates from watchlist status with stronger evidence.
About this signal
Agent Memory Repo — Cognition's open spec for git-backed agent memory is tracked by RepoRadar as an agent project in the Agent Infrastructure section. First seen 2026-10-10; the source record was last checked on 2026-10-10. The current verdict is 'watch' with a Silver tier and moderate setup difficulty. Agent Memory Repo — Cognition's open spec for git-backed agent memory leads on open-source/build quality (8.4) and workflow potential (7.9); its lowest signal is momentum (6.0), so factor that in before investing setup time. This page summarizes the public evidence on the linked source page and states where additional review is still needed.
How this item is evaluated
The Agent Memory Repo — Cognition's open spec for git-backed agent memory record combines a 7.1/10 composite score with separate popularity (100.0), risk (conditional), and setup (moderate) signals. See the scoring methodology for the current weights and evidence definitions.
Putting this into practice? Read How to vet an AI agent or MCP server before you wire it in for the checklist behind this score.
Risk explanation
Agents commit and push memory edits with no human in the loop — keep trial repos local, review Dreaming cleanups, and treat shared remotes as writable agent surface; Memory accumulates personal and team context (people, customers, processes); a configured remote turns that into shared sensitive state — decide what may be remembered before connecting one; Young spec (created 2026-10-04, no releases yet); automatic startup and scheduled Dreaming are not included in the trial; not hands-on tested.