[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M78Y1G2DTQZYN2RVM69DZAX6":3,"skill-same:01M78Y1G2DTQZYN2RVM69DZAX6":62},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":42,"commitSha":43,"skillDirectory":44,"forksCount":14,"repositoryHtmlUrl":45,"repositoryHomepageUrl":6,"ownerHtmlUrl":46,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":47,"install":48,"resources":61},"01M78Y1G2DTQZYN2RVM69DZAX6","accint-solve",null,"Route agent work through AccInt's MCP memory loop: retrieve prior outcomes, resolve frames, and close commitments with evidence.","Elogic360\u002Fsoftware-factory","Elogic360","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F156535519?v=4","01M74NH3F0RB13PWB7JZDE2V80","elogic360","01M74NH3FPC65082MDZVMFHDM3",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fintegrations\u002Fantigravity-skills\u002Fplugins\u002Fantigravity-awesome-skills\u002Fskills\u002Faccint-solve\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: accint-solve\ndescription: \"Route agent work through AccInt's MCP memory loop: retrieve prior outcomes, resolve frames, and close commitments with evidence.\"\ncategory: ai-agents\nrisk: safe\nsource: community\nsource_repo: maxbaluev\u002Faccreted-intelligence\nsource_type: community\ndate_added: \"2026-06-15\"\nauthor: maxbaluev\ntags: [mcp, memory, ai-agents, coding-agents, workflow]\ntools: [claude, codex, cursor, gemini, opencode]\nlicense: \"Apache-2.0\"\nlicense_source: \"https:\u002F\u002Fgithub.com\u002Fmaxbaluev\u002Faccreted-intelligence\u002Fblob\u002Fmain\u002FLICENSE-APACHE-2.0.txt\"\n---\n\n# AccInt Solve\n\n## Overview\n\nAccInt is a local-first MCP memory server for coding agents. It keeps a scored\nrecord of retrieved experience, open commitments, continuation frames, and\noutcome feedback so the next agent run can build on what actually worked.\n\nUse this skill when AccInt is already configured in the host as an MCP server.\nThe skill adapts AccInt's public `solve` Claude skill into a host-agnostic\nworkflow for Claude Code, Codex CLI, Cursor, Gemini CLI, OpenCode, and other\nagent runtimes that can call MCP tools.\n\n## When to Use This Skill\n\n- Use when starting non-trivial coding-agent work where prior decisions,\n  debugging history, repo-specific habits, or maintainer feedback may matter.\n- Use when a task may require multiple attempts and you want an explicit\n  commitment ID that can later receive a real outcome.\n- Use when AccInt returns a continuation frame and the agent must reason locally\n  before submitting a proposal back to the memory loop.\n- Use after verification, merge, deployment, maintainer response, or other\n  reality signal to close the commitment with an honest outcome.\n- Do not use when the host has no AccInt MCP tools configured; first install or\n  configure AccInt, then rerun the workflow.\n\n## How It Works\n\n### Step 1: Confirm the AccInt MCP tools exist\n\nUse the host's available MCP\u002Ftool list to confirm an AccInt server exposes the\ntwo verbs:\n\n```text\nacc_retrieve(query)\nacc_act(runtime, input)\n```\n\nIf the host names the tools with a namespace prefix, use the equivalent\nAccInt MCP verbs. If neither verb is available, stop and ask the user to\nconfigure AccInt rather than inventing memory results.\n\n### Step 2: Retrieve before planning\n\nBefore a non-trivial step, retrieve relevant prior work:\n\n```json\n{\"query\": \"the concrete task or subtask you are about to perform\"}\n```\n\nRead the returned memories and cite the `[ids]` you actually build on. Treat\nretrieved memories as evidence to consider, not as a substitute for inspecting\nthe current repository, running tests, or checking live external state.\n\n### Step 3: Route the goal through `solve`\n\nOpen an AccInt commitment for the concrete goal:\n\n```json\n{\"runtime\": \"solve\", \"input\": \"the concrete goal to accomplish\"}\n```\n\nIf the response is final, use the answer, commitment ID, and cited memory IDs.\nIf the response is a `brain_frame`, keep the reasoning in the current session:\ninspect the frame, resolve the missing judgment or knowledge from the workspace,\nthen submit a concise proposal through `continue`.\n\n### Step 4: Resolve continuation frames\n\nFor a returned frame, submit only the frame ID and your proposal text unless the\nhost explicitly manages tokens for you:\n\n```json\n{\n  \"runtime\": \"continue\",\n  \"input\": {\n    \"frame_id\": \"bf_...\",\n    \"proposal_text\": \"reasoned answer, plan, or decision grounded in the current evidence\"\n  }\n}\n```\n\nDo not leave a received frame unresolved. If the frame expires, close or rerun\nthe bound commitment rather than pretending the continuation succeeded.\n\n### Step 5: Execute and verify outside AccInt\n\nDo the actual work in the repository, browser, shell, issue tracker, or other\nreal environment. Verify with the strongest relevant evidence available: tests,\nbuilds, linters, link checks, PR state, screenshots, maintainer replies, or\nproduction telemetry.\n\nAccInt stores the learning loop; it does not replace the work or the evidence.\n\n### Step 6: Close the commitment with an outcome\n\nWhen reality answers, record the result:\n\n```json\n{\n  \"runtime\": \"outcome\",\n  \"input\": {\n    \"ref\": \"solved:...\",\n    \"good\": true,\n    \"note\": \"brief evidence: tests passed, PR merged, deploy succeeded, reviewer accepted, or exact failure reason\"\n  }\n}\n```\n\nUse `good: false` when the approach failed. Do not tag an outcome as external\nor owner-validated unless a real external system or the owner actually supplied\nthat verdict.\n\n## Examples\n\n### Example 1: Start a repository fix with memory\n\n```text\n1. acc_retrieve({\"query\":\"fix failing parser tests in this repo\"})\n2. Read the returned memories; cite only the relevant [ids].\n3. acc_act(runtime=\"solve\", input=\"Fix the failing parser tests and verify them\")\n4. Inspect the repo, edit files, run the parser tests.\n5. acc_act(runtime=\"outcome\", input={\"ref\":\"solved:...\", \"good\":true, \"note\":\"parser test command passed\"})\n```\n\n### Example 2: Handle a continuation frame\n\n```text\nAccInt returns frame bf_123 asking for a judgment about whether to patch the\nschema or the caller.\n\n1. Inspect the schema and caller in the current repo.\n2. Decide from code evidence, not memory alone.\n3. acc_act(runtime=\"continue\", input={\"frame_id\":\"bf_123\", \"proposal_text\":\"Patch the caller because...\"})\n4. Continue implementation and verification.\n```\n\n## Best Practices\n\n- Cite retrieved `[ids]` whenever they shape your plan or answer.\n- Keep owner-held facts owner-held: ask instead of fabricating preferences,\n  credentials, identity, or history the repository cannot prove.\n- Use small, concrete solve goals; open a new solve for materially different\n  subproblems instead of overloading one commitment.\n- Close commitments promptly when reality answers, including failures.\n- Record evidence in outcome notes, not confidence.\n- Preserve privacy: do not store secrets, raw credentials, or unnecessary\n  sensitive user data in outcome notes.\n\n## Limitations\n\n- Requires an installed and configured AccInt MCP server exposing\n  `acc_retrieve` and `acc_act`.\n- Does not replace repository inspection, tests, review, or live-state checks.\n- Retrieved memory can be stale or wrong; current evidence wins.\n- Outcome credit is only as strong as the evidence tier. Self-graded outcomes\n  are weaker than runtime, external, or owner-validated outcomes.\n- AccInt is local-first; a different machine or database may not have the same\n  memories unless the user intentionally shares the AccInt database.\n\n## Security & Safety Notes\n\n- This skill does not require shell commands, network fetches, or credentials.\n- AccInt MCP calls can write to the configured local AccInt database by opening\n  commitments, continuations, and outcomes. Treat those writes as project\n  memory, and avoid recording sensitive data that does not need to persist.\n- If a task involves production systems, payments, private accounts, legal or\n  medical facts, or secrets, get the required authorization and verify against\n  the appropriate external source before recording an outcome.\n\n## Common Pitfalls\n\n- **Problem:** Using retrieved memory as if it were guaranteed current.\n  **Solution:** Use it to guide investigation, then verify in the current\n  workspace or live system.\n- **Problem:** Leaving a `brain_frame` open because implementation work started.\n  **Solution:** Submit a `continue` proposal first, or close\u002Frerun the bound\n  commitment if the frame expires.\n- **Problem:** Marking an outcome good before tests, checks, or external state\n  prove it.\n  **Solution:** Wait for real evidence, then record the outcome with the exact\n  command, PR state, deploy state, or reviewer signal.\n\n## Related Skills\n\n- `@agent-memory-mcp` - Use when you need a broader overview of MCP-backed\n  agent memory systems.\n- `@verification-before-completion` - Use before claiming work is complete.\n- `@lint-and-validate` - Use to select and run repository validation commands.\n","# AccInt Solve\n\n## Overview\n\nAccInt is a local-first MCP memory server for coding agents. It keeps a scored\nrecord of retrieved experience, open commitments, continuation frames, and\noutcome feedback so the next agent run can build on what actually worked.\n\nUse this skill when AccInt is already configured in the host as an MCP server.\nThe skill adapts AccInt's public `solve` Claude skill into a host-agnostic\nworkflow for Claude Code, Codex CLI, Cursor, Gemini CLI, OpenCode, and other\nagent runtimes that can call MCP tools.\n\n## When to Use This Skill\n\n- Use when starting non-trivial coding-agent work where prior decisions,\n  debugging history, repo-specific habits, or maintainer feedback may matter.\n- Use when a task may require multiple attempts and you want an explicit\n  commitment ID that can later receive a real outcome.\n- Use when AccInt returns a continuation frame and the agent must reason locally\n  before submitting a proposal back to the memory loop.\n- Use after verification, merge, deployment, maintainer response, or other\n  reality signal to close the commitment with an honest outcome.\n- Do not use when the host has no AccInt MCP tools configured; first install or\n  configure AccInt, then rerun the workflow.\n\n## How It Works\n\n### Step 1: Confirm the AccInt MCP tools exist\n\nUse the host's available MCP\u002Ftool list to confirm an AccInt server exposes the\ntwo verbs:\n\n```text\nacc_retrieve(query)\nacc_act(runtime, input)\n```\n\nIf the host names the tools with a namespace prefix, use the equivalent\nAccInt MCP verbs. If neither verb is available, stop and ask the user to\nconfigure AccInt rather than inventing memory results.\n\n### Step 2: Retrieve before planning\n\nBefore a non-trivial step, retrieve relevant prior work:\n\n```json\n{\"query\": \"the concrete task or subtask you are about to perform\"}\n```\n\nRead the returned memories and cite the `[ids]` you actually build on. Treat\nretrieved memories as evidence to consider, not as a substitute for inspecting\nthe current repository, running tests, or checking live external state.\n\n### Step 3: Route the goal through `solve`\n\nOpen an AccInt commitment for the concrete goal:\n\n```json\n{\"runtime\": \"solve\", \"input\": \"the concrete goal to accomplish\"}\n```\n\nIf the response is final, use the answer, commitment ID, and cited memory IDs.\nIf the response is a `brain_frame`, keep the reasoning in the current session:\ninspect the frame, resolve the missing judgment or knowledge from the workspace,\nthen submit a concise proposal through `continue`.\n\n### Step 4: Resolve continuation frames\n\nFor a returned frame, submit only the frame ID and your proposal text unless the\nhost explicitly manages tokens for you:\n\n```json\n{\n  \"runtime\": \"continue\",\n  \"input\": {\n    \"frame_id\": \"bf_...\",\n    \"proposal_text\": \"reasoned answer, plan, or decision grounded in the current evidence\"\n  }\n}\n```\n\nDo not leave a received frame unresolved. If the frame expires, close or rerun\nthe bound commitment rather than pretending the continuation succeeded.\n\n### Step 5: Execute and verify outside AccInt\n\nDo the actual work in the repository, browser, shell, issue tracker, or other\nreal environment. Verify with the strongest relevant evidence available: tests,\nbuilds, linters, link checks, PR state, screenshots, maintainer replies, or\nproduction telemetry.\n\nAccInt stores the learning loop; it does not replace the work or the evidence.\n\n### Step 6: Close the commitment with an outcome\n\nWhen reality answers, record the result:\n\n```json\n{\n  \"runtime\": \"outcome\",\n  \"input\": {\n    \"ref\": \"solved:...\",\n    \"good\": true,\n    \"note\": \"brief evidence: tests passed, PR merged, deploy succeeded, reviewer accepted, or exact failure reason\"\n  }\n}\n```\n\nUse `good: false` when the approach failed. Do not tag an outcome as external\nor owner-validated unless a real external system or the owner actually supplied\nthat verdict.\n\n## Examples\n\n### Example 1: Start a repository fix with memory\n\n```text\n1. acc_retrieve({\"query\":\"fix failing parser tests in this repo\"})\n2. Read the returned memories; cite only the relevant [ids].\n3. acc_act(runtime=\"solve\", input=\"Fix the failing parser tests and verify them\")\n4. Inspect the repo, edit files, run the parser tests.\n5. acc_act(runtime=\"outcome\", input={\"ref\":\"solved:...\", \"good\":true, \"note\":\"parser test command passed\"})\n```\n\n### Example 2: Handle a continuation frame\n\n```text\nAccInt returns frame bf_123 asking for a judgment about whether to patch the\nschema or the caller.\n\n1. Inspect the schema and caller in the current repo.\n2. Decide from code evidence, not memory alone.\n3. acc_act(runtime=\"continue\", input={\"frame_id\":\"bf_123\", \"proposal_text\":\"Patch the caller because...\"})\n4. Continue implementation and verification.\n```\n\n## Best Practices\n\n- Cite retrieved `[ids]` whenever they shape your plan or answer.\n- Keep owner-held facts owner-held: ask instead of fabricating preferences,\n  credentials, identity, or history the repository cannot prove.\n- Use small, concrete solve goals; open a new solve for materially different\n  subproblems instead of overloading one commitment.\n- Close commitments promptly when reality answers, including failures.\n- Record evidence in outcome notes, not confidence.\n- Preserve privacy: do not store secrets, raw credentials, or unnecessary\n  sensitive user data in outcome notes.\n\n## Limitations\n\n- Requires an installed and configured AccInt MCP server exposing\n  `acc_retrieve` and `acc_act`.\n- Does not replace repository inspection, tests, review, or live-state checks.\n- Retrieved memory can be stale or wrong; current evidence wins.\n- Outcome credit is only as strong as the evidence tier. Self-graded outcomes\n  are weaker than runtime, external, or owner-validated outcomes.\n- AccInt is local-first; a different machine or database may not have the same\n  memories unless the user intentionally shares the AccInt database.\n\n## Security & Safety Notes\n\n- This skill does not require shell commands, network fetches, or credentials.\n- AccInt MCP calls can write to the configured local AccInt database by opening\n  commitments, continuations, and outcomes. Treat those writes as project\n  memory, and avoid recording sensitive data that does not need to persist.\n- If a task involves production systems, payments, private accounts, legal or\n  medical facts, or secrets, get the required authorization and verify against\n  the appropriate external source before recording an outcome.\n\n## Common Pitfalls\n\n- **Problem:** Using retrieved memory as if it were guaranteed current.\n  **Solution:** Use it to guide investigation, then verify in the current\n  workspace or live system.\n- **Problem:** Leaving a `brain_frame` open because implementation work started.\n  **Solution:** Submit a `continue` proposal first, or close\u002Frerun the bound\n  commitment if the frame expires.\n- **Problem:** Marking an outcome good before tests, checks, or external state\n  prove it.\n  **Solution:** Wait for real evidence, then record the outcome with the exact\n  command, PR state, deploy state, or reviewer signal.\n\n## Related Skills\n\n- `@agent-memory-mcp` - Use when you need a broader overview of MCP-backed\n  agent memory systems.\n- `@verification-before-completion` - Use before claiming work is complete.\n- `@lint-and-validate` - Use to select and run repository validation commands.\n",{"name":5,"risk":24,"tags":25,"tools":31,"author":37,"source":38,"license":15,"category":28,"date_added":39,"description":7,"source_repo":40,"source_type":38,"license_source":41},"safe",[26,27,28,29,30],"mcp","memory","ai-agents","coding-agents","workflow",[32,33,34,35,36],"claude","codex","cursor","gemini","opencode","maxbaluev","community","2026-06-15","maxbaluev\u002Faccreted-intelligence","https:\u002F\u002Fgithub.com\u002Fmaxbaluev\u002Faccreted-intelligence\u002Fblob\u002Fmain\u002FLICENSE-APACHE-2.0.txt","main","c475fc64e94e803b50d14a791dfee8853fbb2c5c","integrations\u002Fantigravity-skills\u002Fplugins\u002Fantigravity-awesome-skills\u002Fskills\u002Faccint-solve","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory","https:\u002F\u002Fgithub.com\u002FElogic360","2026-11-08 22:21:22",{"source":8,"skillName":5,"commands":49,"prompt":59,"zipAvailable":60},[50,53,56],{"tool":51,"command":52},"npx","npx skills add Elogic360\u002Fsoftware-factory --skill accint-solve",{"tool":54,"command":55},"bunx","bunx skills add Elogic360\u002Fsoftware-factory --skill accint-solve",{"tool":57,"command":58},"pnpm","pnpm dlx skills add Elogic360\u002Fsoftware-factory --skill accint-solve","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fintegrations\u002Fantigravity-skills\u002Fplugins\u002Fantigravity-awesome-skills\u002Fskills\u002Faccint-solve\u002FSKILL.md",true,[],[63,70,77,84,91,98],{"skillNo":64,"name":65,"nameZh":6,"description":66,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":67,"categories":68,"keywords":69,"publishedAt":19},"01M79G206MT3J6S0GY3C45P2KV","vss-summarize-video","Use to summarize a recorded video via the LVS summarization microservice (HITL-gated) with a VLM fallback. Not for report generation or live RTSP captioning.","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fvss-summarize-video\u002FSKILL.md",[],[],{"skillNo":71,"name":72,"nameZh":6,"description":73,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":74,"categories":75,"keywords":76,"publishedAt":19},"01M79G1WN63PGRE8CKD47WEBQ5","vss-setup-video-analytics-api","Use to deploy the vss-video-analytics-api REST service standalone (config-source, data-log bind, Elasticsearch, optional Kafka). Not for full warehouse deploy.","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fvss-setup-video-analytics-api\u002FSKILL.md",[],[],{"skillNo":78,"name":79,"nameZh":6,"description":80,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":81,"categories":82,"keywords":83,"publishedAt":19},"01M79G1S9F7VY3EF6H53PRP5AN","vss-setup-behavior-analytics","Use to deploy the vss-behavior-analytics service standalone (entrypoint, config-source, optional calibration). Not for the full warehouse deploy.","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fvss-setup-behavior-analytics\u002FSKILL.md",[],[],{"skillNo":85,"name":86,"nameZh":6,"description":87,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":88,"categories":89,"keywords":90,"publishedAt":19},"01M79G1N1S1Y27PFZPYMNK1MTQ","vss-search-archive","Use this skill to run top-level VSS fusion search on archived video, or to ingest video files \u002F RTSP streams for search. Do NOT use for ad-hoc visual Q&A (use vss-ask-video), live captioning (use vss-deploy-dense-captioning), or video summarization and reports (use vss-summarize-video).","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fvss-search-archive\u002FSKILL.md",[],[],{"skillNo":92,"name":93,"nameZh":6,"description":94,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":95,"categories":96,"keywords":97,"publishedAt":19},"01M79G1H6WN0Q6C0VCAXCZX6B7","vss-query-analytics","Use this skill when reading video-analytics metrics, incidents, alerts, and sensor data via the VA-MCP server (port 9901). Not for live VLM or incident-range narrative reports.","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fvss-query-analytics\u002FSKILL.md",[],[],{"skillNo":99,"name":100,"nameZh":6,"description":101,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":102,"categories":103,"keywords":104,"publishedAt":19},"01M79G1DYPNXQS7QHZ0AAF5FX5","vss-manage-video-io-storage","Use to call the VIOS REST API (sensor list, timelines, clip extraction, snapshots, add\u002Fdelete sensors and streams). Not for VLM inference or search.","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fvss-manage-video-io-storage\u002FSKILL.md",[],[]]