{ "schemaVersion": "1.0", "item": { "slug": "deeprecall", "name": "DeepRecall", "source": "tencent", "type": "skill", "category": "AI 智能", "sourceUrl": "https://clawhub.ai/Stefan27-4/deeprecall", "canonicalUrl": "https://clawhub.ai/Stefan27-4/deeprecall", "targetPlatform": "OpenClaw" }, "install": { "downloadMode": "redirect", "downloadUrl": "/downloads/deeprecall", "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=deeprecall", "sourcePlatform": "tencent", "targetPlatform": "OpenClaw", "installMethod": "Manual import", "extraction": "Extract archive", "prerequisites": [ "OpenClaw" ], "packageFormat": "ZIP package", "includedAssets": [ "deep_recall.py", "memory_indexer.py", "memory_scanner.py", "model_pairs.py", "provider_bridge.py", "SKILL.md" ], "primaryDoc": "SKILL.md", "quickSetup": [ "Download the package from Yavira.", "Extract the archive and review SKILL.md first.", "Import or place the package into your OpenClaw setup." ], "agentAssist": { "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.", "steps": [ "Download the package from Yavira.", "Extract it into a folder your agent can access.", "Paste one of the prompts below and point your agent at the extracted folder." ], "prompts": [ { "label": "New install", "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete." }, { "label": "Upgrade existing", "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run." } ] }, "sourceHealth": { "source": "tencent", "status": "healthy", "reason": "direct_download_ok", "recommendedAction": "download", "checkedAt": "2026-04-30T16:55:25.780Z", "expiresAt": "2026-05-07T16:55:25.780Z", "httpStatus": 200, "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=network", "contentType": "application/zip", "probeMethod": "head", "details": { "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=network", "contentDisposition": "attachment; filename=\"network-1.0.0.zip\"", "redirectLocation": null, "bodySnippet": null }, "scope": "source", "summary": "Source download looks usable.", "detail": "Yavira can redirect you to the upstream package for this source.", "primaryActionLabel": "Download for OpenClaw", "primaryActionHref": "/downloads/deeprecall" }, "validation": { "installChecklist": [ "Use the Yavira download entry.", "Review SKILL.md after the package is downloaded.", "Confirm the extracted package contains the expected setup assets." ], "postInstallChecks": [ "Confirm the extracted package includes the expected docs or setup files.", "Validate the skill or prompts are available in your target agent workspace.", "Capture any manual follow-up steps the agent could not complete." ] }, "downloadPageUrl": "https://openagent3.xyz/downloads/deeprecall", "agentPageUrl": "https://openagent3.xyz/skills/deeprecall/agent", "manifestUrl": "https://openagent3.xyz/skills/deeprecall/agent.json", "briefUrl": "https://openagent3.xyz/skills/deeprecall/agent.md" }, "agentAssist": { "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.", "steps": [ "Download the package from Yavira.", "Extract it into a folder your agent can access.", "Paste one of the prompts below and point your agent at the extracted folder." ], "prompts": [ { "label": "New install", "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete." }, { "label": "Upgrade existing", "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run." } ] }, "documentation": { "source": "clawhub", "primaryDoc": "SKILL.md", "sections": [ { "title": "DeepRecall v2 — OpenClaw Skill", "body": "Pure-Python recursive memory for persistent AI agents.\nImplements the Anamnesis Architecture: \"The soul stays small, the mind scales forever.\"" }, { "title": "Description", "body": "DeepRecall gives AI agents infinite memory by recursively querying their own\nmemory files through a manager→workers→synthesis RLM loop — entirely in Python.\nNo Deno runtime, no fast-rlm subprocess, no vector database. Just markdown files\nand HTTP calls to any OpenAI-compatible LLM endpoint.\n\nWhen the agent needs to recall something, DeepRecall:\n\nScans the workspace for memory files (scoped by category)\nIndexes file metadata — headers, topics, dates, people\nManager selects the most relevant files from the index\nWorkers (parallel) extract exact verbatim quotes from each file\nSynthesis combines quotes into a cited, grounded answer\n\nWorkers are constrained by anti-hallucination prompts to return only verbatim\nquotes. The synthesis step cites every claim with (filename:line)." }, { "title": "Installation", "body": "pip install deep-recall\n\nOr install from source:\n\ngit clone https://github.com/Stefan27-4/DeepRecall\ncd DeepRecall && pip install ." }, { "title": "Dependencies", "body": "httpx (preferred) or requests — HTTP client for LLM calls\nPyYAML — config parsing\nPython ≥ 3.10\nAn LLM provider configured in OpenClaw\n\nv2 breaking change: Deno and fast-rlm are no longer required.\nThe entire RLM loop runs in-process as pure Python." }, { "title": "Quick Start", "body": "from deep_recall import recall\n\nresult = recall(\"What did we decide about the project architecture?\")\nprint(result)" }, { "title": "recall(query, scope, workspace, verbose, config_overrides) → str", "body": "The primary entry point. Runs the full manager→workers→synthesis loop.\n\nfrom deep_recall import recall\n\nresult = recall(\n \"Find all mentions of budget discussions\",\n scope=\"memory\", # \"memory\" | \"identity\" | \"project\" | \"all\"\n verbose=True, # print progress to stdout\n config_overrides={\n \"max_files\": 5, # max files the manager can select\n },\n)\n\nParameterTypeDefaultDescriptionquerystr(required)What to recall / search forscopestr\"memory\"File scope — see ScopesworkspacePath | Noneauto-detectOverride workspace pathverboseboolFalsePrint provider, model, file selection infoconfig_overridesdict | NoneNoneOverride max_files and other settings\n\nReturns: A string containing the recalled information with source citations,\nor a [DeepRecall] status message if no files/results were found." }, { "title": "recall_quick(query, verbose) → str", "body": "Fast, cheap recall scoped to identity files. Best for simple lookups.\n\nfrom deep_recall import recall_quick\n\nname = recall_quick(\"What is my human's name?\")\n\nEquivalent to recall(query, scope=\"identity\", config_overrides={\"max_files\": 2})." }, { "title": "recall_deep(query, verbose) → str", "body": "Thorough recall across all workspace files. Best for cross-referencing.\n\nfrom deep_recall import recall_deep\n\nsummary = recall_deep(\"Summarize all decisions from March\")\n\nEquivalent to recall(query, scope=\"all\", config_overrides={\"max_files\": 5})." }, { "title": "CLI", "body": "python deep_recall.py [scope]\n\n# Examples\npython deep_recall.py \"What was the first project we worked on?\"\npython deep_recall.py \"Find budget discussions\" all" }, { "title": "Scopes", "body": "Scopes control which files DeepRecall searches. Narrower scopes are faster\nand cheaper.\n\nScopeFiles IncludedSpeedCostUse CaseidentitySOUL.md, IDENTITY.md, MEMORY.md, USER.md, TOOLS.md, HEARTBEAT.md, AGENTS.md⚡ FastestCheapest\"What's my name?\"memoryIdentity files + memory/LONG_TERM.md + memory/*.md daily logs🔄 FastLow\"What did we do last week?\"projectAll readable workspace files (skips binaries, node_modules, .git)🐢 SlowerMedium\"Find that config change\"allIdentity + memory + project (everything)🐌 SlowestHighest\"Search everything\"" }, { "title": "File Categories", "body": "DeepRecall classifies discovered files into categories:\n\nsoul — SOUL.md, IDENTITY.md — who the agent IS (always in context)\nmind — MEMORY.md, USER.md, TOOLS.md, HEARTBEAT.md, AGENTS.md — compact orientation\nlong-term — memory/LONG_TERM.md — full detailed memories, grows forever\ndaily-log — memory/YYYY-MM-DD.md — raw daily logs\nworkspace — everything else (project files, configs, docs)" }, { "title": "Configuration", "body": "DeepRecall reads your existing OpenClaw setup — no additional config files needed." }, { "title": "Provider Resolution", "body": "Provider, API key, and model are resolved automatically from:\n\n~/.openclaw/openclaw.json — primary model setting\n~/.openclaw/agents/main/agent/models.json — provider base URLs\n~/.openclaw/credentials/ — cached tokens (e.g. GitHub Copilot)\nEnvironment variables — fallback (ANTHROPIC_API_KEY, OPENAI_API_KEY, GOOGLE_API_KEY, etc. (18+ providers supported, all optional))" }, { "title": "Supported Providers (20+)", "body": "Anthropic · OpenAI · Google (Gemini) · GitHub Copilot · OpenRouter · Ollama ·\nDeepSeek · Mistral · Together · Groq · Fireworks · Cohere · Perplexity ·\nSambaNova · Cerebras · xAI · Minimax · Zhipu (GLM) · Moonshot (Kimi) · Qwen" }, { "title": "Auto Model Pairing", "body": "The manager and synthesis steps use your primary model. Workers use a cheaper\nsub-agent model automatically:\n\nPrimary ModelWorker ModelClaude Opus 4 / 4.6Claude Sonnet 4Claude Sonnet 4 / 4.5Claude Haiku 3.5GPT-4o / GPT-4GPT-4o-miniGemini 2.5 ProGemini 2.0 FlashDeepSeek ReasonerDeepSeek ChatLlama 3.1 70BLlama 3.1 8B" }, { "title": "config_overrides", "body": "Pass overrides via the config_overrides parameter:\n\nrecall(\"query\", config_overrides={\n \"max_files\": 5, # max files manager can select (default: 3)\n})" }, { "title": "Skill Files", "body": "FilePurposedeep_recall.pyPublic API — recall, recall_quick, recall_deep, RLM loopprovider_bridge.pyResolves LLM provider, API key, base URL from OpenClaw configmodel_pairs.pyMaps primary models to cheaper worker modelsmemory_scanner.pyDiscovers and categorises workspace files by scopememory_indexer.pyBuilds a structured Memory Index (topics, people, timeline)__init__.pyPackage exports" }, { "title": "Memory Layout", "body": "Recommended workspace structure for the Anamnesis Architecture:\n\n~/.openclaw/workspace/\n├── SOUL.md # Identity — always in context, never grows\n├── IDENTITY.md # Core agent facts\n├── MEMORY.md # Compact index (~100 lines), auto-loaded each session\n├── USER.md # About the human\n├── AGENTS.md # Agent behavior rules\n├── TOOLS.md # Tool-specific notes\n└── memory/\n ├── LONG_TERM.md # Full memories — grows forever, searched via DeepRecall\n ├── 2026-03-05.md # Daily raw log\n ├── 2026-03-04.md\n └── ..." }, { "title": "⚠️ Privacy Notice", "body": "DeepRecall reads your workspace memory files and sends their contents to your configured LLM provider (Anthropic, OpenAI, Gemini, etc.) to perform recall. This is how it works — there is no local-only mode.\n\nWhat gets sent:\n\nFile metadata (names, headings, topics) → to the manager LLM\nFull file contents of selected files → to worker LLMs\nThis may include personal notes, daily logs, project files\n\nWhat is NOT sent:\n\nAPI keys and credentials (read locally for auth, never in prompts)\nFiles outside your workspace\n\nCredentials used locally:\n\n~/.openclaw/openclaw.json and ~/.openclaw/credentials/* — to resolve your LLM provider\nEnv vars (ANTHROPIC_API_KEY, OPENAI_API_KEY, GOOGLE_API_KEY, etc. (18+ providers supported, all optional)) — as fallback if no OpenClaw config found" }, { "title": "Recommended Memory Architecture", "body": "DeepRecall works best with a two-tier memory system:" }, { "title": "Tier 1: MEMORY.md (The Index)", "body": "Auto-loaded every session — keep it SMALL (~100 lines)\nContains: quick reference facts, active projects, key metrics, a table of contents pointing to LONG_TERM.md sections\nThink of it as your orientation file — \"what do I need to know right now?\"\nIf it grows past ~120 lines, move details to LONG_TERM.md" }, { "title": "Tier 2: memory/LONG_TERM.md (The Encyclopedia)", "body": "Never auto-loaded — searched via DeepRecall when needed\nContains: full context, decisions with reasoning, timestamps, bugs with fixes, architecture details\nGrows forever — never delete, only append\nThe devil is in the details — \"the diary entry, not the Wikipedia summary\"" }, { "title": "Tier 3: memory/YYYY-MM-DD.md (Daily Logs)", "body": "Raw notes of what happened each day\nDistilled into LONG_TERM.md at end of day" }, { "title": "Nightly Sync Routine", "body": "At the end of each day (or via cron/heartbeat):\n\nRead today's daily log\nAppend key events, decisions, lessons, bugs, metrics to LONG_TERM.md\nUpdate MEMORY.md index table if new topics appeared\n\n⚠️ Setting this up? Ask your human before restructuring existing memory files.\nShow them this recommendation and let them decide how to organize their agent's memory." }, { "title": "License", "body": "MIT — see LICENSE." } ], "body": "DeepRecall v2 — OpenClaw Skill\n\nPure-Python recursive memory for persistent AI agents. Implements the Anamnesis Architecture: \"The soul stays small, the mind scales forever.\"\n\nDescription\n\nDeepRecall gives AI agents infinite memory by recursively querying their own memory files through a manager→workers→synthesis RLM loop — entirely in Python. No Deno runtime, no fast-rlm subprocess, no vector database. Just markdown files and HTTP calls to any OpenAI-compatible LLM endpoint.\n\nWhen the agent needs to recall something, DeepRecall:\n\nScans the workspace for memory files (scoped by category)\nIndexes file metadata — headers, topics, dates, people\nManager selects the most relevant files from the index\nWorkers (parallel) extract exact verbatim quotes from each file\nSynthesis combines quotes into a cited, grounded answer\n\nWorkers are constrained by anti-hallucination prompts to return only verbatim quotes. The synthesis step cites every claim with (filename:line).\n\nInstallation\npip install deep-recall\n\n\nOr install from source:\n\ngit clone https://github.com/Stefan27-4/DeepRecall\ncd DeepRecall && pip install .\n\nDependencies\nhttpx (preferred) or requests — HTTP client for LLM calls\nPyYAML — config parsing\nPython ≥ 3.10\nAn LLM provider configured in OpenClaw\n\nv2 breaking change: Deno and fast-rlm are no longer required. The entire RLM loop runs in-process as pure Python.\n\nQuick Start\nfrom deep_recall import recall\n\nresult = recall(\"What did we decide about the project architecture?\")\nprint(result)\n\nAPI\nrecall(query, scope, workspace, verbose, config_overrides) → str\n\nThe primary entry point. Runs the full manager→workers→synthesis loop.\n\nfrom deep_recall import recall\n\nresult = recall(\n \"Find all mentions of budget discussions\",\n scope=\"memory\", # \"memory\" | \"identity\" | \"project\" | \"all\"\n verbose=True, # print progress to stdout\n config_overrides={\n \"max_files\": 5, # max files the manager can select\n },\n)\n\nParameter\tType\tDefault\tDescription\nquery\tstr\t(required)\tWhat to recall / search for\nscope\tstr\t\"memory\"\tFile scope — see Scopes\nworkspace\tPath | None\tauto-detect\tOverride workspace path\nverbose\tbool\tFalse\tPrint provider, model, file selection info\nconfig_overrides\tdict | None\tNone\tOverride max_files and other settings\n\nReturns: A string containing the recalled information with source citations, or a [DeepRecall] status message if no files/results were found.\n\nrecall_quick(query, verbose) → str\n\nFast, cheap recall scoped to identity files. Best for simple lookups.\n\nfrom deep_recall import recall_quick\n\nname = recall_quick(\"What is my human's name?\")\n\n\nEquivalent to recall(query, scope=\"identity\", config_overrides={\"max_files\": 2}).\n\nrecall_deep(query, verbose) → str\n\nThorough recall across all workspace files. Best for cross-referencing.\n\nfrom deep_recall import recall_deep\n\nsummary = recall_deep(\"Summarize all decisions from March\")\n\n\nEquivalent to recall(query, scope=\"all\", config_overrides={\"max_files\": 5}).\n\nCLI\npython deep_recall.py [scope]\n\n# Examples\npython deep_recall.py \"What was the first project we worked on?\"\npython deep_recall.py \"Find budget discussions\" all\n\nScopes\n\nScopes control which files DeepRecall searches. Narrower scopes are faster and cheaper.\n\nScope\tFiles Included\tSpeed\tCost\tUse Case\nidentity\tSOUL.md, IDENTITY.md, MEMORY.md, USER.md, TOOLS.md, HEARTBEAT.md, AGENTS.md\t⚡ Fastest\tCheapest\t\"What's my name?\"\nmemory\tIdentity files + memory/LONG_TERM.md + memory/*.md daily logs\t🔄 Fast\tLow\t\"What did we do last week?\"\nproject\tAll readable workspace files (skips binaries, node_modules, .git)\t🐢 Slower\tMedium\t\"Find that config change\"\nall\tIdentity + memory + project (everything)\t🐌 Slowest\tHighest\t\"Search everything\"\nFile Categories\n\nDeepRecall classifies discovered files into categories:\n\nsoul — SOUL.md, IDENTITY.md — who the agent IS (always in context)\nmind — MEMORY.md, USER.md, TOOLS.md, HEARTBEAT.md, AGENTS.md — compact orientation\nlong-term — memory/LONG_TERM.md — full detailed memories, grows forever\ndaily-log — memory/YYYY-MM-DD.md — raw daily logs\nworkspace — everything else (project files, configs, docs)\nConfiguration\n\nDeepRecall reads your existing OpenClaw setup — no additional config files needed.\n\nProvider Resolution\n\nProvider, API key, and model are resolved automatically from:\n\n~/.openclaw/openclaw.json — primary model setting\n~/.openclaw/agents/main/agent/models.json — provider base URLs\n~/.openclaw/credentials/ — cached tokens (e.g. GitHub Copilot)\nEnvironment variables — fallback (ANTHROPIC_API_KEY, OPENAI_API_KEY, GOOGLE_API_KEY, etc. (18+ providers supported, all optional))\nSupported Providers (20+)\n\nAnthropic · OpenAI · Google (Gemini) · GitHub Copilot · OpenRouter · Ollama · DeepSeek · Mistral · Together · Groq · Fireworks · Cohere · Perplexity · SambaNova · Cerebras · xAI · Minimax · Zhipu (GLM) · Moonshot (Kimi) · Qwen\n\nAuto Model Pairing\n\nThe manager and synthesis steps use your primary model. Workers use a cheaper sub-agent model automatically:\n\nPrimary Model\tWorker Model\nClaude Opus 4 / 4.6\tClaude Sonnet 4\nClaude Sonnet 4 / 4.5\tClaude Haiku 3.5\nGPT-4o / GPT-4\tGPT-4o-mini\nGemini 2.5 Pro\tGemini 2.0 Flash\nDeepSeek Reasoner\tDeepSeek Chat\nLlama 3.1 70B\tLlama 3.1 8B\nconfig_overrides\n\nPass overrides via the config_overrides parameter:\n\nrecall(\"query\", config_overrides={\n \"max_files\": 5, # max files manager can select (default: 3)\n})\n\nSkill Files\nFile\tPurpose\ndeep_recall.py\tPublic API — recall, recall_quick, recall_deep, RLM loop\nprovider_bridge.py\tResolves LLM provider, API key, base URL from OpenClaw config\nmodel_pairs.py\tMaps primary models to cheaper worker models\nmemory_scanner.py\tDiscovers and categorises workspace files by scope\nmemory_indexer.py\tBuilds a structured Memory Index (topics, people, timeline)\n__init__.py\tPackage exports\nMemory Layout\n\nRecommended workspace structure for the Anamnesis Architecture:\n\n~/.openclaw/workspace/\n├── SOUL.md # Identity — always in context, never grows\n├── IDENTITY.md # Core agent facts\n├── MEMORY.md # Compact index (~100 lines), auto-loaded each session\n├── USER.md # About the human\n├── AGENTS.md # Agent behavior rules\n├── TOOLS.md # Tool-specific notes\n└── memory/\n ├── LONG_TERM.md # Full memories — grows forever, searched via DeepRecall\n ├── 2026-03-05.md # Daily raw log\n ├── 2026-03-04.md\n └── ...\n\n⚠️ Privacy Notice\n\nDeepRecall reads your workspace memory files and sends their contents to your configured LLM provider (Anthropic, OpenAI, Gemini, etc.) to perform recall. This is how it works — there is no local-only mode.\n\nWhat gets sent:\n\nFile metadata (names, headings, topics) → to the manager LLM\nFull file contents of selected files → to worker LLMs\nThis may include personal notes, daily logs, project files\n\nWhat is NOT sent:\n\nAPI keys and credentials (read locally for auth, never in prompts)\nFiles outside your workspace\n\nCredentials used locally:\n\n~/.openclaw/openclaw.json and ~/.openclaw/credentials/* — to resolve your LLM provider\nEnv vars (ANTHROPIC_API_KEY, OPENAI_API_KEY, GOOGLE_API_KEY, etc. (18+ providers supported, all optional)) — as fallback if no OpenClaw config found\nRecommended Memory Architecture\n\nDeepRecall works best with a two-tier memory system:\n\nTier 1: MEMORY.md (The Index)\nAuto-loaded every session — keep it SMALL (~100 lines)\nContains: quick reference facts, active projects, key metrics, a table of contents pointing to LONG_TERM.md sections\nThink of it as your orientation file — \"what do I need to know right now?\"\nIf it grows past ~120 lines, move details to LONG_TERM.md\nTier 2: memory/LONG_TERM.md (The Encyclopedia)\nNever auto-loaded — searched via DeepRecall when needed\nContains: full context, decisions with reasoning, timestamps, bugs with fixes, architecture details\nGrows forever — never delete, only append\nThe devil is in the details — \"the diary entry, not the Wikipedia summary\"\nTier 3: memory/YYYY-MM-DD.md (Daily Logs)\nRaw notes of what happened each day\nDistilled into LONG_TERM.md at end of day\nNightly Sync Routine\n\nAt the end of each day (or via cron/heartbeat):\n\nRead today's daily log\nAppend key events, decisions, lessons, bugs, metrics to LONG_TERM.md\nUpdate MEMORY.md index table if new topics appeared\n\n⚠️ Setting this up? Ask your human before restructuring existing memory files. Show them this recommendation and let them decide how to organize their agent's memory.\n\nLicense\n\nMIT — see LICENSE." }, "trust": { "sourceLabel": "tencent", "provenanceUrl": "https://clawhub.ai/Stefan27-4/deeprecall", "publisherUrl": "https://clawhub.ai/Stefan27-4/deeprecall", "owner": "Stefan27-4", "version": "1.0.8", "license": null, "verificationStatus": "Indexed source record" }, "links": { "detailUrl": "https://openagent3.xyz/skills/deeprecall", "downloadUrl": "https://openagent3.xyz/downloads/deeprecall", "agentUrl": "https://openagent3.xyz/skills/deeprecall/agent", "manifestUrl": "https://openagent3.xyz/skills/deeprecall/agent.json", "briefUrl": "https://openagent3.xyz/skills/deeprecall/agent.md" } }