Tree Ring Memory
Framework-agnostic agent memory lifecycle skill and Rust-native CLI for explicit recall, evidence-backed memory, forgetting, audit, consolidation, DOX/Revolve sync, and local SQLite/FTS storage.
Open the source and read safety notes before installing.
Citation facts
Source-backed facts for citing this resource, derived directly from the registry — also available as plain text for AI assistants.
- Canonical URL
- https://heyclau.de/entry/skills/tree-ring-memory
- Source URLs
- https://github.com/TerminallyLazy/Tree-Ring-Memory/blob/main/docs/integrations/agent-skill.md, https://github.com/TerminallyLazy/Tree-Ring-Memory, https://terminallylazy.github.io/Tree-Ring-Memory/
- Safety notes
- The repository includes an optional shell installer; review `install.sh` before running it in locked-down or production environments., Memory writes are explicit through commands such as `remember`, `evidence`, `import`, consolidation, maintenance, TUI actions, or deliberate agent calls; do not configure agents to store everything by default., Run `tree-ring dox sync` and `tree-ring revolve sync` with `--dry-run` first because adapters summarize source artifacts into memory events., Use `forget`, redaction, supersession, audit, and maintenance commands when a memory is wrong, sensitive, stale, or no longer useful., Tree Ring Memory is in protocol-preview status, so verify command behavior and integration expectations before relying on it for critical workflows.
- Privacy notes
- Project memory normally lives in a local `.tree-ring/` SQLite store and generated guidance files; treat that directory as project data., Tree Ring Memory does not scrape transcripts or run a hidden background recorder, but agent-mediated prompts can still include sensitive context if the active agent chooses to write it., DOX/Revolve adapters and integration scans inspect local source roots to create concise, source-linked summaries; preview outputs before committing them to durable memory., Do not store secrets, credentials, private keys, raw chain-of-thought, or regulated personal data; prefer short redacted summaries with source references.
- Platform compatibility
- claude-code (native-skill), codex (native-skill), windsurf (native-skill), gemini (native-skill), cursor (adapter), cli (manual-context)
- Author
- TerminallyLazy
- Submitted by
- TerminallyLazy
- Claim status
- unclaimed
- Last verified
- 2026-07-08
Safety notes
- The repository includes an optional shell installer; review `install.sh` before running it in locked-down or production environments.
- Memory writes are explicit through commands such as `remember`, `evidence`, `import`, consolidation, maintenance, TUI actions, or deliberate agent calls; do not configure agents to store everything by default.
- Run `tree-ring dox sync` and `tree-ring revolve sync` with `--dry-run` first because adapters summarize source artifacts into memory events.
- Use `forget`, redaction, supersession, audit, and maintenance commands when a memory is wrong, sensitive, stale, or no longer useful.
- Tree Ring Memory is in protocol-preview status, so verify command behavior and integration expectations before relying on it for critical workflows.
Privacy notes
- Project memory normally lives in a local `.tree-ring/` SQLite store and generated guidance files; treat that directory as project data.
- Tree Ring Memory does not scrape transcripts or run a hidden background recorder, but agent-mediated prompts can still include sensitive context if the active agent chooses to write it.
- DOX/Revolve adapters and integration scans inspect local source roots to create concise, source-linked summaries; preview outputs before committing them to durable memory.
- Do not store secrets, credentials, private keys, raw chain-of-thought, or regulated personal data; prefer short redacted summaries with source references.
Prerequisites
- Rust/Cargo for source installs unless using a future verified release archive path.
- A project or agent workflow where durable memory should be explicit, scoped, and auditable.
- A local memory root such as `.tree-ring/` for project-scoped stores, skill guidance, and CLI references.
- Review of generated `.tree-ring/SKILL.md`, `.tree-ring/AGENTS.md`, and `.tree-ring/CLI.md` before wiring them into an agent host.
Schema details
- Install type
- package
- Reading time
- 8 min
- Difficulty score
- 76
- Troubleshooting
- Yes
- Breaking changes
- No
- Scope
- Source repo
- Skill type
- capability-pack
- Skill level
- expert
- Verification
- validated
- Verified at
- 2026-07-08
| Platform | Support | Install path |
|---|---|---|
| claude-code | Native | .claude/skills/<skill-name>/SKILL.md |
| codex | Native | .agents/skills/<skill-name>/SKILL.md |
| windsurf | Native | .windsurf/skills/<skill-name>/SKILL.md |
| gemini | Native | .gemini/skills/<skill-name>/SKILL.md or .agents/skills/<skill-name>/SKILL.md |
| cursor | Adapter | .cursor/rules/<skill-name>.mdc |
| cli | Manual | AGENTS.md or tool-specific context file |
Full copyable content
# Install the Rust-native CLI from source with Cargo
cargo install --git https://github.com/TerminallyLazy/Tree-Ring-Memory tree-ring-memory-cli --locked
# Initialize local project memory and try recall
tree-ring init
tree-ring remember "Use project-scoped recall before risky release changes." --event-type lesson --scope project
tree-ring recall "release changes"
tree-ring evidence "The eval passed after the fix." --outcome promoted --evidence-ref evals/run-042
tree-ring tuiAbout this resource
Tree Ring Memory
Tree Ring Memory is a framework-agnostic, local-first memory lifecycle layer for AI agents. It helps agents remember useful decisions, warnings, preferences, and evidence without turning memory into raw transcript storage.
The checked-in skill teaches an agent when to recall, remember, redact, forget, and avoid memory capture. The Rust-native CLI owns the local store, recall, evidence capture, audit, consolidation, import/export, maintenance, terminal onboarding, and Ratatui operator console.
Knowledge Freshness
This listing was verified on 2026-07-08 against the public repository,
README, portable SKILL.md, integration guide, press kit, and Rust-native CLI
launch article. Tree Ring Memory is currently marked as protocol-preview, so
re-check the release and CLI help before production integration.
Retrieval Sources
- https://github.com/TerminallyLazy/Tree-Ring-Memory
- https://github.com/TerminallyLazy/Tree-Ring-Memory/blob/main/README.md
- https://github.com/TerminallyLazy/Tree-Ring-Memory/blob/main/skills/tree-ring-memory/SKILL.md
- https://github.com/TerminallyLazy/Tree-Ring-Memory/blob/main/docs/integrations/agent-skill.md
- https://terminallylazy.github.io/Tree-Ring-Memory/press-kit.md
- https://terminallylazy.github.io/Tree-Ring-Memory/launch/rust-native-agent-memory-cli.md
Core Workflow
Install the CLI from source with Cargo:
cargo install --git https://github.com/TerminallyLazy/Tree-Ring-Memory tree-ring-memory-cli --locked
Initialize local memory and try recall:
tree-ring init
tree-ring remember "Use project-scoped recall before risky release changes." --event-type lesson --scope project
tree-ring recall "release changes"
tree-ring evidence "The eval passed after the fix." --outcome promoted --evidence-ref evals/run-042
tree-ring tui
Use adapter previews when source artifacts already exist:
tree-ring dox sync --source-root . --dry-run
tree-ring revolve sync --source-root revolve --dry-run
tree-ring integrations scan --source-root .
Capability Scope
- Project-scoped and global memory recall.
- Explicit memory capture through
remember. - Evidence-backed outcomes through
tree-ring evidence. - Ring lifecycle model: cambium, outer, inner, heartwood, scar, and seed.
- Local SQLite/FTS storage and JSONL import/export.
- Forgetting, redaction, supersession, audit, consolidation, and maintenance.
- DOX and Revolve source adapters with dry-run previews.
- Portable
SKILL.md, generated.tree-ring/guidance, and agent-framework discovery. - Terminal onboarding and Ratatui operator console.
Compatibility
Native
- Rust CLI / terminal workflows: use
tree-ringcommands directly. - Agent hosts that support local skills or instruction packs: load
skills/tree-ring-memory/SKILL.mdor the generated.tree-ring/SKILL.md.
Manual Adaptation
- Claude Code, Codex, Gemini, OpenCode, and Generic AGENTS workflows: point
project bridge files at
.tree-ring/SKILL.md,.tree-ring/AGENTS.md, and.tree-ring/CLI.mdso host-specific instructions stay scoped to the current project. - DOX-aware repositories: merge the generated
.tree-ring/AGENTS.mdguidance manually when agents should discover Tree Ring rules before entering the memory directory.
Production Rules
- Treat memory as a lifecycle, not a transcript dump.
- Recall before changing architecture, storage, security, privacy, release behavior, or project conventions.
- Store concise lessons, decisions, warnings, user preferences, or future seeds only when they will materially help future work.
- Prefer
tree-ring evidencefor evaluated outcomes, incidents, checkpoints, or reviewed run artifacts. - Run source adapters with
--dry-runfirst and preserve source references. - Do not promote weak claims to heartwood; use outer or seed until evidence is stronger.
- Redact or delete memories that are wrong, sensitive, stale, or superseded.
- Keep source documents authoritative; memory points back to them instead of replacing them.
Troubleshooting
Issue: tree-ring is not on PATH after a global install
Fix: open a new shell, run $HOME/.local/bin/tree-ring, or add the install
bin directory to the current terminal path.
Issue: An agent wants to store too much context Fix: apply the portable skill guidance and store only durable, concise, privacy-safe lessons with source references.
Issue: A source adapter produces noisy memories
Fix: keep the --dry-run output, narrow the source root, and write only the
summaries that are useful and source-linked.
Output Contract
When using this skill, produce:
- Relevant recalled lessons, warnings, preferences, or decisions.
- A clear decision on whether new durable memory is warranted.
- The exact
tree-ringcommand to write, evidence-link, redact, delete, or supersede memory. - Safety and privacy notes for any proposed write.
- Source references for memories derived from docs, issues, PRs, evals, DOX, or Revolve artifacts.
Duplicate Review
Checked content/skills, generated catalog text, and current source entries
for Tree Ring Memory, local-first agent memory lifecycle, DOX/Revolve memory
adapters, and Rust-native agent memory CLI listings. Existing entries cover
general agent skills, plugin creation, portability, and other AI memory
resources, but no HeyClaude skills entry lists Tree Ring Memory's portable
skill plus Rust-native local memory lifecycle CLI.
Editorial Disclosure
Submitted by the project owner as a source-backed community listing. The project is MIT-licensed, protocol-preview software; review the installer, generated guidance files, and local privacy boundary before adopting it in a production agent workflow.
Source citations
Add this badge to your README
How it compares
Tree Ring Memory side by side with 3 alternatives on trust, install, platform support, and disclosed safety notes — all from reviewed registry metadata.
3 trust signals differ across this comparison (Package trust, Source provenance, Submitter).
Next steps differ across entries — use the actions in the table below to copy install commands and source links per resource.
| Field | Framework-agnostic agent memory lifecycle skill and Rust-native CLI for explicit recall, evidence-backed memory, forgetting, audit, consolidation, DOX/Revolve sync, and local SQLite/FTS storage. Open dossier | Expert agent skills cross-platform adapter capability pack for porting Claude Code SKILL.md workflows to Codex, Cursor, Windsurf, and Generic AGENTS runbooks while preserving scope, safety notes, and source-backed contracts. Open dossier | MIT-licensed BrowserAct Agent Skill pack for installing and operating the `browser-act` browser automation CLI from Claude Code, Codex, OpenClaw, Cursor, OpenCode, Windsurf, Gemini CLI, and other skills-compatible agents. Open dossier | Expert code-review capability pack for deterministic PR audits, risk-ranked findings, and low-noise fix planning without SaaS lock-in. Open dossier |
|---|---|---|---|---|
| Next stepsDiffers | ||||
| Trust | ||||
| Review status | ReviewedMaintainer reviewed | ReviewedMaintainer reviewed | ReviewedMaintainer reviewed | ReviewedMaintainer reviewed |
| Package trustDiffers | Package not verified | Package not verified | Package not verified | Package verified2026-04-11 |
| Source provenanceDiffers | Source-backed | Submission linkedSource submission | Source-backed | No submission link |
| SubmitterDiffers | TerminallyLazy | kiannidev | — | — |
| Install risk | Review first | Review first | Review first | Low risk |
| Notes | Safety ✓ Privacy ✓ | Safety ✓ Privacy ✓ | Safety ✓ Privacy ✓ | Safety ✓ Privacy ✓ |
| Brand | — | — | — | |
| Category | skills | skills | skills | skills |
| Source | source-backed | source-backed | source-backed | first-party |
| Author | TerminallyLazy | kiannidev | BrowserAct | JSONbored |
| Added | 2026-07-08 | 2026-06-14 | 2026-06-18 | 2026-04-11 |
| Platforms | Claude CodeCodexWindsurfGeminiCursorCLI | Claude CodeCodexWindsurfGeminiCursorCLI | Claude CodeCodexWindsurfGeminiCursorCLIVS Code | Claude CodeCodexWindsurfGeminiCursorCLI |
| Source repo | — | — | — | — |
| Safety notes | ✓The repository includes an optional shell installer; review `install.sh` before running it in locked-down or production environments. Memory writes are explicit through commands such as `remember`, `evidence`, `import`, consolidation, maintenance, TUI actions, or deliberate agent calls; do not configure agents to store everything by default. Run `tree-ring dox sync` and `tree-ring revolve sync` with `--dry-run` first because adapters summarize source artifacts into memory events. Use `forget`, redaction, supersession, audit, and maintenance commands when a memory is wrong, sensitive, stale, or no longer useful. Tree Ring Memory is in protocol-preview status, so verify command behavior and integration expectations before relying on it for critical workflows. | ✓Not every Claude Code command or MCP tool exists on other platforms; do not imply feature parity without verification. Adapted skills must preserve destructive-action warnings and approval requirements from the source workflow. Cross-platform copies should not strip `safetyNotes` or `privacyNotes` to save space. Platform-specific secret handling differs; never copy Claude Code env patterns blindly into other hosts. This skill produces adaptation plans; it must not run destructive commands on the target platform without explicit approval. | ✓BrowserAct can open pages, click, type, upload files, inspect state, capture screenshots, read page text, handle dialogs, export cookies, capture network requests, and operate logged-in browser sessions. Use BrowserAct only on sites, accounts, and data sources where the user has authorization. Do not use it to evade access controls, violate site terms, scrape disallowed data, or bypass rate limits. The entry skill declares confirmation gates for browser creation, deletion, local Chrome profile import, proxy/security changes, logins, form submissions, file uploads, and other sensitive operations; preserve those gates in agent workflows. `solve-captcha` may send the challenge image to BrowserAct's verification-assistance service according to the skill metadata; do not use it with sensitive or unauthorized pages. `remote-assist` can generate a live handoff URL for a human to take over. Treat that URL as access to the active browser session. Skill Forge can generate reusable automation skills from explored sites. Review generated scripts, selectors, network assumptions, output schemas, and site authorization before reusing them at scale. | ✓Installs from a downloaded package and may run local commands or scaffold files as part of the workflow; review the package and any generated changes before applying. |
| Privacy notes | ✓Project memory normally lives in a local `.tree-ring/` SQLite store and generated guidance files; treat that directory as project data. Tree Ring Memory does not scrape transcripts or run a hidden background recorder, but agent-mediated prompts can still include sensitive context if the active agent chooses to write it. DOX/Revolve adapters and integration scans inspect local source roots to create concise, source-linked summaries; preview outputs before committing them to durable memory. Do not store secrets, credentials, private keys, raw chain-of-thought, or regulated personal data; prefer short redacted summaries with source references. | ✓Adaptation reviews may expose internal runbooks, customer workflows, and repository conventions embedded in source skills. Generic AGENTS exports shared across teams can leak path patterns and tool names if pasted into public issues. Public adaptation summaries should describe platform gaps and preserved boundaries, not full internal skill bodies. | ✓BrowserAct workflows can expose page content, screenshots, URLs, credentials typed into forms, cookies, browser profiles, uploaded files, downloaded files, network requests, HAR data, session names, browser descriptions, and logs. The BrowserAct skill metadata states that cookies, login sessions, page content, credentials, and browser profile data stay local, except the CAPTCHA challenge image when `solve-captcha` is invoked. Chrome-direct and profile import workflows can connect agents to existing local browser state. Treat those modes as account access, not a blank test browser. Log reports, feedback, Discord support, generated Skill Forge packages, and shared screenshots can leak private browsing or account context if submitted without review. Managed proxy, stealth browser, and API-key features create additional BrowserAct service dependencies beyond local CLI execution. | ✓Operates on your local project files and any context you share in the session; review what you expose before sharing — nothing is sent beyond the model unless a step calls an external service. |
| Prerequisites |
|
|
|
|
| Install | | — | | |
| Config | — | — | — | — |
| Citations | ||||
| Claim | Unclaimed | Unclaimed | Unclaimed | Unclaimed |
Related guides
Source-backed guides for putting this to work.
Adding Plugin Install Hints to Developer CLIs
Write a claude-code-hint tag to stderr under CLAUDECODE, and Claude Code prompts users once to install your official-marketplace plugin.
Signals
Loading live community signals…
A short, calm digest of reviewed Claude resources. Unsubscribe any time.