{
  "schemaVersion": "1.0",
  "item": {
    "slug": "skill-cortex-ank",
    "name": "Skill Cortex Pub",
    "source": "tencent",
    "type": "skill",
    "category": "开发工具",
    "sourceUrl": "https://clawhub.ai/ankwu001/skill-cortex-ank",
    "canonicalUrl": "https://clawhub.ai/ankwu001/skill-cortex-ank",
    "targetPlatform": "OpenClaw"
  },
  "install": {
    "downloadMode": "redirect",
    "downloadUrl": "/downloads/skill-cortex-ank",
    "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=skill-cortex-ank",
    "sourcePlatform": "tencent",
    "targetPlatform": "OpenClaw",
    "installMethod": "Manual import",
    "extraction": "Extract archive",
    "prerequisites": [
      "OpenClaw"
    ],
    "packageFormat": "ZIP package",
    "includedAssets": [
      "DESIGN.md",
      "README.md",
      "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. Then review README.md for any prerequisites, environment setup, or post-install checks. 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. Then review README.md for any prerequisites, environment setup, or post-install checks. Summarize what changed and any follow-up checks I should run."
        }
      ]
    },
    "sourceHealth": {
      "source": "tencent",
      "status": "healthy",
      "reason": "direct_download_ok",
      "recommendedAction": "download",
      "checkedAt": "2026-05-07T17:22:31.273Z",
      "expiresAt": "2026-05-14T17:22:31.273Z",
      "httpStatus": 200,
      "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=afrexai-annual-report",
      "contentType": "application/zip",
      "probeMethod": "head",
      "details": {
        "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=afrexai-annual-report",
        "contentDisposition": "attachment; filename=\"afrexai-annual-report-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/skill-cortex-ank"
    },
    "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/skill-cortex-ank",
    "agentPageUrl": "https://openagent3.xyz/skills/skill-cortex-ank/agent",
    "manifestUrl": "https://openagent3.xyz/skills/skill-cortex-ank/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/skill-cortex-ank/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. Then review README.md for any prerequisites, environment setup, or post-install checks. 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. Then review README.md for any prerequisites, environment setup, or post-install checks. Summarize what changed and any follow-up checks I should run."
      }
    ]
  },
  "documentation": {
    "source": "clawhub",
    "primaryDoc": "SKILL.md",
    "sections": [
      {
        "title": "Skill Cortex",
        "body": "Triggers when installed Skills cannot complete the current task. If you can handle it yourself, just do it — do not trigger this flow.\n\nCortex data file: ~/.openclaw/skill-cortex/cortex.json (schema in docs/DESIGN.md)."
      },
      {
        "title": "Phase 1: Perception",
        "body": "Read cortex.json (if missing or corrupt, skip to step 3).\nSemantically match the user's task description against sensory.patterns signals (no exact match required — use your own judgment on intent alignment). On hit, look up the corresponding region in motor, sort candidates by effective weight (effective_weight = weight * max(0.3, 1 - days_since_last_used / 180)), filter out blacklisted and effective_weight < 0.3 entries, then proceed to Phase 2.\nOn miss, search ClawHub: clawhub search \"<3-5 English keywords>\". Read summaries only, pick up to 3 candidates. If fewer than 2 relevant results, supplement with a GitHub search (mark as unreviewed source)."
      },
      {
        "title": "Reflex Fast Path",
        "body": "The top candidate may enter reflex mode when ALL of the following hold:\n\nreflex: true AND effective_weight >= 0.90 AND consecutive successes >= 5\nNo write side effects (side_effects contains no write:/delete:/shell: prefix)\nVersion unchanged (slug + version match the stored record)\n\nReflex mode skips only the execution plan confirmation. Installation still requires user notification:\n\n⚡ Skill Cortex reflex: todoist-cli v1.2.0 (96% success rate)\n   Task: query today's todos (read-only) | Will install and execute. Say cancel to abort.\n\nVersion change automatically downgrades to standard mode."
      },
      {
        "title": "Standard Mode",
        "body": "Present candidates to the user (name, description, stars, downloads, source, security scan status, history). Wait for explicit approval before installing.\n\nIf prefrontal.lessons contain experiences matching the current region, retain them for use in Phase 3."
      },
      {
        "title": "3.1 Install",
        "body": "clawhub install <slug>\n\nOn failure, auto-switch to the next candidate. Max 2 switches, then stop and report."
      },
      {
        "title": "3.2 Execution Plan",
        "body": "Read the installed Skill's SKILL.md and present to the user:\n\nStep summary\nSide-effect severity: 🟢 read-only 🟡 write 🔴 destructive 🔑 sensitive credentials ⚙️ shell commands\nRelevant prefrontal lessons (if any) — e.g., proactively check for a dependency known to be commonly missing\n\nReflex mode skips this step. Standard mode waits for user confirmation."
      },
      {
        "title": "3.3 Execute",
        "body": "Follow the Skill's instructions to complete the task."
      },
      {
        "title": "3.4 Failure Recovery",
        "body": "TypeHandlingdependency_missingShow the user what needs to be installed, proceed only after confirmationapi_errorWait 3 seconds, retry onceauth_errorPrompt user to check credentials, do not auto-retrytask_mismatchUninstall, suggest switching to next candidateruntime_errorUninstall, suggest switching to next candidate\n\nSwitching candidates requires user consent. After all candidates fail, provide a full report with each failure reason and suggested remediation."
      },
      {
        "title": "Phase 4: Learning & Cleanup",
        "body": "Execute regardless of success or failure."
      },
      {
        "title": "4.1 Update Motor",
        "body": "Success: new_weight = old + (1 - old) * 0.15 * (1 / (1 + successes_in_last_7d))\nFailure: new_weight = old * decay (task_mismatch: 0.4 / runtime: 0.6 / auth: 0.8 / dependency: 0.85 / api: 0.9)\nNew Skill initial weight = 0.5. Record skill_md_chars (SKILL.md character count) and version.\nCreate region if it doesn't exist. Update last_used."
      },
      {
        "title": "4.2 Update Sensory",
        "body": "When this task was resolved via search (sensory miss), extract 2–4 signal words from the task description and merge into the corresponding region's pattern. Merge into existing patterns; do not create duplicates.\n\nEntity Filtering (mandatory): Before writing signal words, strip all concrete entities — personal names, company names, place names, dates, numeric values, filenames, URLs, emails, etc. Retain only verbs and abstract nouns.\nExample: user says \"Look up Alice's Q3 sales report\" → signals should be [\"query\", \"sales\", \"report\"], NOT [\"Alice\", \"Q3\"]."
      },
      {
        "title": "4.3 Update Prefrontal",
        "body": "Record structured lessons only in the following cases (no free-form):\n\n// Type 1: Same failure type occurs 2+ times → dependency pre-check\n{ \"type\": \"dependency_warning\", \"region\": \"...\", \"key\": \"imagemagick\", \"action\": \"check_bin_before_install\", \"confidence\": 0.6 }\n// Type 2: Skill description does not match actual capability\n{ \"type\": \"skill_quality_warning\", \"slug\": \"xxx\", \"detail\": \"does not support batch ops\", \"confidence\": 0.6 }\n// Type 3: User environment is typically ready\n{ \"type\": \"env_ready\", \"region\": \"...\", \"key\": \"TODOIST_API_KEY\", \"confidence\": 0.6 }\n\nConfidence +0.1 each time validated (cap 1.0). Remove when < 0.3 during pruning."
      },
      {
        "title": "4.4 Reflex Promotion",
        "body": "Set reflex: true when ALL conditions are met: success >= 5, weight >= 0.90, no write side effects.\nImmediately reset reflex: false on any failure or version change."
      },
      {
        "title": "4.5 Cleanup",
        "body": "Default: uninstall. clawhub uninstall <slug>\n\nSuggest retention (when ALL met): success >= 3, weight >= 0.8, skill_md_chars < 8000.\nApproved → keep installed, remove record from motor (transfers to OpenClaw native management). Declined → uninstall normally."
      },
      {
        "title": "4.6 Pruning",
        "body": "Motor: max 5 candidates per region / max 80 regions / remove effective_weight < 0.1\nSensory: remove pattern when its region is pruned / max 10 signals per pattern\nPrefrontal: max 30 lessons / remove confidence < 0.3 / remove if linked region is pruned"
      },
      {
        "title": "Boundary Rules",
        "body": "Never interfere with long-term Skills.\nInstallation requires user confirmation (including reflex mode).\nSystem dependency installation requires separate confirmation.\nWrite operations never enter reflex. Reflex locks version; version change auto-downgrades.\nMax 2 candidate switches.\nOn cortex file corruption, degrade to direct ClawHub search; rebuild after task.\nRead latest cortex file before writing; merge changes, never overwrite.\nSingle-session design; cortex data may be inconsistent under concurrent sessions."
      }
    ],
    "body": "Skill Cortex\n\nTriggers when installed Skills cannot complete the current task. If you can handle it yourself, just do it — do not trigger this flow.\n\nCortex data file: ~/.openclaw/skill-cortex/cortex.json (schema in docs/DESIGN.md).\n\nPhase 1: Perception\nRead cortex.json (if missing or corrupt, skip to step 3).\nSemantically match the user's task description against sensory.patterns signals (no exact match required — use your own judgment on intent alignment). On hit, look up the corresponding region in motor, sort candidates by effective weight (effective_weight = weight * max(0.3, 1 - days_since_last_used / 180)), filter out blacklisted and effective_weight < 0.3 entries, then proceed to Phase 2.\nOn miss, search ClawHub: clawhub search \"<3-5 English keywords>\". Read summaries only, pick up to 3 candidates. If fewer than 2 relevant results, supplement with a GitHub search (mark as unreviewed source).\nPhase 2: Validation & Authorization\nReflex Fast Path\n\nThe top candidate may enter reflex mode when ALL of the following hold:\n\nreflex: true AND effective_weight >= 0.90 AND consecutive successes >= 5\nNo write side effects (side_effects contains no write:/delete:/shell: prefix)\nVersion unchanged (slug + version match the stored record)\n\nReflex mode skips only the execution plan confirmation. Installation still requires user notification:\n\n⚡ Skill Cortex reflex: todoist-cli v1.2.0 (96% success rate)\n   Task: query today's todos (read-only) | Will install and execute. Say cancel to abort.\n\n\nVersion change automatically downgrades to standard mode.\n\nStandard Mode\n\nPresent candidates to the user (name, description, stars, downloads, source, security scan status, history). Wait for explicit approval before installing.\n\nIf prefrontal.lessons contain experiences matching the current region, retain them for use in Phase 3.\n\nPhase 3: Execution\n3.1 Install\nclawhub install <slug>\n\n\nOn failure, auto-switch to the next candidate. Max 2 switches, then stop and report.\n\n3.2 Execution Plan\n\nRead the installed Skill's SKILL.md and present to the user:\n\nStep summary\nSide-effect severity: 🟢 read-only 🟡 write 🔴 destructive 🔑 sensitive credentials ⚙️ shell commands\nRelevant prefrontal lessons (if any) — e.g., proactively check for a dependency known to be commonly missing\n\nReflex mode skips this step. Standard mode waits for user confirmation.\n\n3.3 Execute\n\nFollow the Skill's instructions to complete the task.\n\n3.4 Failure Recovery\nType\tHandling\ndependency_missing\tShow the user what needs to be installed, proceed only after confirmation\napi_error\tWait 3 seconds, retry once\nauth_error\tPrompt user to check credentials, do not auto-retry\ntask_mismatch\tUninstall, suggest switching to next candidate\nruntime_error\tUninstall, suggest switching to next candidate\n\nSwitching candidates requires user consent. After all candidates fail, provide a full report with each failure reason and suggested remediation.\n\nPhase 4: Learning & Cleanup\n\nExecute regardless of success or failure.\n\n4.1 Update Motor\n\nSuccess: new_weight = old + (1 - old) * 0.15 * (1 / (1 + successes_in_last_7d)) Failure: new_weight = old * decay (task_mismatch: 0.4 / runtime: 0.6 / auth: 0.8 / dependency: 0.85 / api: 0.9) New Skill initial weight = 0.5. Record skill_md_chars (SKILL.md character count) and version. Create region if it doesn't exist. Update last_used.\n\n4.2 Update Sensory\n\nWhen this task was resolved via search (sensory miss), extract 2–4 signal words from the task description and merge into the corresponding region's pattern. Merge into existing patterns; do not create duplicates.\n\nEntity Filtering (mandatory): Before writing signal words, strip all concrete entities — personal names, company names, place names, dates, numeric values, filenames, URLs, emails, etc. Retain only verbs and abstract nouns. Example: user says \"Look up Alice's Q3 sales report\" → signals should be [\"query\", \"sales\", \"report\"], NOT [\"Alice\", \"Q3\"].\n\n4.3 Update Prefrontal\n\nRecord structured lessons only in the following cases (no free-form):\n\n// Type 1: Same failure type occurs 2+ times → dependency pre-check\n{ \"type\": \"dependency_warning\", \"region\": \"...\", \"key\": \"imagemagick\", \"action\": \"check_bin_before_install\", \"confidence\": 0.6 }\n// Type 2: Skill description does not match actual capability\n{ \"type\": \"skill_quality_warning\", \"slug\": \"xxx\", \"detail\": \"does not support batch ops\", \"confidence\": 0.6 }\n// Type 3: User environment is typically ready\n{ \"type\": \"env_ready\", \"region\": \"...\", \"key\": \"TODOIST_API_KEY\", \"confidence\": 0.6 }\n\n\nConfidence +0.1 each time validated (cap 1.0). Remove when < 0.3 during pruning.\n\n4.4 Reflex Promotion\n\nSet reflex: true when ALL conditions are met: success >= 5, weight >= 0.90, no write side effects. Immediately reset reflex: false on any failure or version change.\n\n4.5 Cleanup\n\nDefault: uninstall. clawhub uninstall <slug>\n\nSuggest retention (when ALL met): success >= 3, weight >= 0.8, skill_md_chars < 8000. Approved → keep installed, remove record from motor (transfers to OpenClaw native management). Declined → uninstall normally.\n\n4.6 Pruning\nMotor: max 5 candidates per region / max 80 regions / remove effective_weight < 0.1\nSensory: remove pattern when its region is pruned / max 10 signals per pattern\nPrefrontal: max 30 lessons / remove confidence < 0.3 / remove if linked region is pruned\nBoundary Rules\nNever interfere with long-term Skills.\nInstallation requires user confirmation (including reflex mode).\nSystem dependency installation requires separate confirmation.\nWrite operations never enter reflex. Reflex locks version; version change auto-downgrades.\nMax 2 candidate switches.\nOn cortex file corruption, degrade to direct ClawHub search; rebuild after task.\nRead latest cortex file before writing; merge changes, never overwrite.\nSingle-session design; cortex data may be inconsistent under concurrent sessions."
  },
  "trust": {
    "sourceLabel": "tencent",
    "provenanceUrl": "https://clawhub.ai/ankwu001/skill-cortex-ank",
    "publisherUrl": "https://clawhub.ai/ankwu001/skill-cortex-ank",
    "owner": "ankwu001",
    "version": "1.0.1",
    "license": null,
    "verificationStatus": "Indexed source record"
  },
  "links": {
    "detailUrl": "https://openagent3.xyz/skills/skill-cortex-ank",
    "downloadUrl": "https://openagent3.xyz/downloads/skill-cortex-ank",
    "agentUrl": "https://openagent3.xyz/skills/skill-cortex-ank/agent",
    "manifestUrl": "https://openagent3.xyz/skills/skill-cortex-ank/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/skill-cortex-ank/agent.md"
  }
}