[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M736WN1AFBZHG71M63RG4XDH":3,"skill-same:01M736WN1AFBZHG71M63RG4XDH":44},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":23,"commitSha":24,"skillDirectory":25,"forksCount":13,"repositoryHtmlUrl":26,"repositoryHomepageUrl":6,"ownerHtmlUrl":27,"ownerBlogUrl":6,"ownerDescription":28,"updatedAt":29,"install":30,"resources":43},"01M736WN1AFBZHG71M63RG4XDH","execution-capsule",null,"Prepare a private execution capsule with a reviewable script and direct or agent-supervised run commands; use only for an already authorized local operation.","graysurf\u002Fagent-runtime-kit","graysurf","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F10785178?v=4","01M736PQHYKHFMNBYF8BT93NM4","01M736PQJWYH7G046WAN7EQCHD",0,"MIT","https:\u002F\u002Fgithub.com\u002Fgraysurf\u002Fagent-runtime-kit\u002Fblob\u002Fmain\u002Ftests\u002Fgolden\u002Fhermes\u002Fplugins\u002Fmeta\u002Fskills\u002Fexecution-capsule\u002Fexpected\u002FSKILL.md",[],[],"2026-11-11 15:11:14",false,"---\nname: execution-capsule\ndescription: >\n  Prepare a private execution capsule with a reviewable script and direct or\n  agent-supervised run commands; use only for an already authorized local\n  operation.\n---\n\n# Execution Capsule\n\n## Contract\n\nPrereqs:\n\n- The underlying local operation is already authorized. Capsule preparation\n  must not infer authorization for destructive scope, provider mutation,\n  credential access, host access, or a materially different task.\n- Resolve and read `core\u002Fpolicies\u002Fexecution-capsules.md` from the selected\n  agent-docs home before constructing any files. That policy remains the\n  canonical safety and lifecycle contract.\n- `agent-out` is available. The target working directory and every allowed\n  path are exact absolute paths.\n- For `workspace`, the workspace capsule must be outside the canonical `cwd`,\n  and every canonical allowed path must remain under that `cwd`. If the\n  `agent-out` allocation violates that containment boundary, stop; do not\n  relocate it manually or switch to `host`.\n- The supervised alternative requires a `codex-cli agent run` implementation\n  whose help exposes `--capsule`. If it is unavailable, still preserve the\n  direct operator path and report that supervised execution is unavailable;\n  do not substitute a raw `codex` invocation.\n\nInputs:\n\n- The authorized operation, exact `cwd`, access class (`workspace` by default),\n  allowed paths, relevant preconditions, validation commands, and a short\n  capsule topic.\n- Explicit operator intent before selecting `host`; a constrained agent never\n  selects or launches host access merely to bypass its environment.\n\nOutputs:\n\n- One private capsule directory allocated outside the repository and, for\n  `workspace`, outside its canonical `cwd`, with:\n\n  ```text\n  \u003Ccapsule>\u002F                 0700\n    manifest.json            0600\n    run.sh                   0700\n  ```\n\n- The absolute capsule path and applicable one-line commands:\n\n  ```sh\n  bash \u003Ccapsule>\u002Frun.sh\n  codex-cli agent run --capsule \u003Ccapsule>\n  ```\n\n- For an explicitly prepared host capsule, the reason host access is required\n  and this operator-authorized command:\n\n  ```sh\n  codex-cli agent run --capsule \u003Ccapsule> --allow-host-access\n  ```\n\nFailure modes:\n\n- The operation, target, authorization, preconditions, or validation boundary\n  is ambiguous.\n- The capsule would contain a secret, token, private key, auth payload, or an\n  unauthorized network\u002Fprovider effect.\n- The capsule path is inside the repository, permissions are not owner-only,\n  an object is a symlink\u002Fhardlink, or the entrypoint digest does not match.\n- A workspace capsule equals or descends from canonical `cwd`, or an allowed\n  path escapes that `cwd`.\n- Host access is needed but the operator has not explicitly selected that\n  route.\n\n## Entrypoint\n\nAllocate a private project-scoped artifact directory:\n\n```sh\nagent-out project --topic \u003Ctopic> --mkdir\n```\n\nConstruct `run.sh` and `manifest.json` there with the active environment's\napproved file-edit mechanism. Follow the manifest schema, digest fields,\nprecondition rules, and validation arrays in\n`core\u002Fpolicies\u002Fexecution-capsules.md`; do not invent a parallel capsule format.\n\n## Workflow\n\n1. Resolve the exact operation, `cwd`, allowed paths, authorization, and access\n   class. Default to `workspace`.\n2. Load `core\u002Fpolicies\u002Fexecution-capsules.md` from the selected docs home and\n   keep its direct\u002Fsupervised parity, governance, and receipt boundaries.\n3. Allocate the capsule through `agent-out project --topic \u003Ctopic> --mkdir`.\n   Never place it in the repository or repo-local `.\u002Fagent-out`. Before\n   authoring, canonicalize the result and enforce the workspace `cwd` and\n   allowed-path containment rules; stop on a mismatch.\n4. Author `run.sh` so it:\n   - starts with `set -euo pipefail`;\n   - enters or verifies the exact target;\n   - checks branch, HEAD, files, or other relevant state before mutation;\n   - uses project-governed commands without disabling hooks or signing;\n   - is idempotent where practical and fails clearly otherwise;\n   - performs essential validation for the direct operator path; and\n   - omits sensitive values and unauthorized provider\u002Fnetwork mutations.\n5. Author `manifest.json` with\n   `schema_version: \"execution-capsule.v1\"`, absolute `cwd`,\n   `entrypoint: \"run.sh\"`, the final script digest, access and allowed paths,\n   optional exact Git preconditions, and validation argv arrays.\n6. Set directory and script mode `0700`, JSON mode `0600`, then recheck object\n   type, ownership, digest, paths, and absence of sensitive values. Verify\n   `codex-cli agent run --help` exposes `--capsule` before claiming the\n   supervised route is currently runnable.\n7. Return the absolute path and one-line commands. Do not make the operator\n   reconstruct a long command from prose.\n8. Do not run `run.sh` directly as the agent. Run a workspace capsule through\n   `codex-cli agent run` only when the user also asked for supervised\n   execution. For host capsules, prepare and report the\n   `--allow-host-access` command for the operator instead of self-launching it.\n9. After supervised execution, report the receipt path and concise final\n   status. Keep receipts and event logs outside project history.\n\n## Boundary\n\nThis skill is the discoverable preparation workflow. It does not replace or\nduplicate the canonical execution-capsule policy, change the nils-cli schema,\ngrant authority, bypass sandbox limits, or turn host access into an agent-side\nescape hatch.\n","# Execution Capsule\n\n## Contract\n\nPrereqs:\n\n- The underlying local operation is already authorized. Capsule preparation\n  must not infer authorization for destructive scope, provider mutation,\n  credential access, host access, or a materially different task.\n- Resolve and read `core\u002Fpolicies\u002Fexecution-capsules.md` from the selected\n  agent-docs home before constructing any files. That policy remains the\n  canonical safety and lifecycle contract.\n- `agent-out` is available. The target working directory and every allowed\n  path are exact absolute paths.\n- For `workspace`, the workspace capsule must be outside the canonical `cwd`,\n  and every canonical allowed path must remain under that `cwd`. If the\n  `agent-out` allocation violates that containment boundary, stop; do not\n  relocate it manually or switch to `host`.\n- The supervised alternative requires a `codex-cli agent run` implementation\n  whose help exposes `--capsule`. If it is unavailable, still preserve the\n  direct operator path and report that supervised execution is unavailable;\n  do not substitute a raw `codex` invocation.\n\nInputs:\n\n- The authorized operation, exact `cwd`, access class (`workspace` by default),\n  allowed paths, relevant preconditions, validation commands, and a short\n  capsule topic.\n- Explicit operator intent before selecting `host`; a constrained agent never\n  selects or launches host access merely to bypass its environment.\n\nOutputs:\n\n- One private capsule directory allocated outside the repository and, for\n  `workspace`, outside its canonical `cwd`, with:\n\n  ```text\n  \u003Ccapsule>\u002F                 0700\n    manifest.json            0600\n    run.sh                   0700\n  ```\n\n- The absolute capsule path and applicable one-line commands:\n\n  ```sh\n  bash \u003Ccapsule>\u002Frun.sh\n  codex-cli agent run --capsule \u003Ccapsule>\n  ```\n\n- For an explicitly prepared host capsule, the reason host access is required\n  and this operator-authorized command:\n\n  ```sh\n  codex-cli agent run --capsule \u003Ccapsule> --allow-host-access\n  ```\n\nFailure modes:\n\n- The operation, target, authorization, preconditions, or validation boundary\n  is ambiguous.\n- The capsule would contain a secret, token, private key, auth payload, or an\n  unauthorized network\u002Fprovider effect.\n- The capsule path is inside the repository, permissions are not owner-only,\n  an object is a symlink\u002Fhardlink, or the entrypoint digest does not match.\n- A workspace capsule equals or descends from canonical `cwd`, or an allowed\n  path escapes that `cwd`.\n- Host access is needed but the operator has not explicitly selected that\n  route.\n\n## Entrypoint\n\nAllocate a private project-scoped artifact directory:\n\n```sh\nagent-out project --topic \u003Ctopic> --mkdir\n```\n\nConstruct `run.sh` and `manifest.json` there with the active environment's\napproved file-edit mechanism. Follow the manifest schema, digest fields,\nprecondition rules, and validation arrays in\n`core\u002Fpolicies\u002Fexecution-capsules.md`; do not invent a parallel capsule format.\n\n## Workflow\n\n1. Resolve the exact operation, `cwd`, allowed paths, authorization, and access\n   class. Default to `workspace`.\n2. Load `core\u002Fpolicies\u002Fexecution-capsules.md` from the selected docs home and\n   keep its direct\u002Fsupervised parity, governance, and receipt boundaries.\n3. Allocate the capsule through `agent-out project --topic \u003Ctopic> --mkdir`.\n   Never place it in the repository or repo-local `.\u002Fagent-out`. Before\n   authoring, canonicalize the result and enforce the workspace `cwd` and\n   allowed-path containment rules; stop on a mismatch.\n4. Author `run.sh` so it:\n   - starts with `set -euo pipefail`;\n   - enters or verifies the exact target;\n   - checks branch, HEAD, files, or other relevant state before mutation;\n   - uses project-governed commands without disabling hooks or signing;\n   - is idempotent where practical and fails clearly otherwise;\n   - performs essential validation for the direct operator path; and\n   - omits sensitive values and unauthorized provider\u002Fnetwork mutations.\n5. Author `manifest.json` with\n   `schema_version: \"execution-capsule.v1\"`, absolute `cwd`,\n   `entrypoint: \"run.sh\"`, the final script digest, access and allowed paths,\n   optional exact Git preconditions, and validation argv arrays.\n6. Set directory and script mode `0700`, JSON mode `0600`, then recheck object\n   type, ownership, digest, paths, and absence of sensitive values. Verify\n   `codex-cli agent run --help` exposes `--capsule` before claiming the\n   supervised route is currently runnable.\n7. Return the absolute path and one-line commands. Do not make the operator\n   reconstruct a long command from prose.\n8. Do not run `run.sh` directly as the agent. Run a workspace capsule through\n   `codex-cli agent run` only when the user also asked for supervised\n   execution. For host capsules, prepare and report the\n   `--allow-host-access` command for the operator instead of self-launching it.\n9. After supervised execution, report the receipt path and concise final\n   status. Keep receipts and event logs outside project history.\n\n## Boundary\n\nThis skill is the discoverable preparation workflow. It does not replace or\nduplicate the canonical execution-capsule policy, change the nils-cli schema,\ngrant authority, bypass sandbox limits, or turn host access into an agent-side\nescape hatch.\n",{"name":5,"description":7},"main","da7119f2ebece510d6955cea72743069f5496c73","tests\u002Fgolden\u002Fhermes\u002Fplugins\u002Fmeta\u002Fskills\u002Fexecution-capsule\u002Fexpected","https:\u002F\u002Fgithub.com\u002Fgraysurf\u002Fagent-runtime-kit","https:\u002F\u002Fgithub.com\u002Fgraysurf","Building reliable tools for teams of AI coding agents · Rust \u002F TypeScript · @sympoies","2026-08-29 05:27:47",{"source":8,"skillName":5,"commands":31,"prompt":41,"zipAvailable":42},[32,35,38],{"tool":33,"command":34},"npx","npx skills add graysurf\u002Fagent-runtime-kit --skill execution-capsule",{"tool":36,"command":37},"bunx","bunx skills add graysurf\u002Fagent-runtime-kit --skill execution-capsule",{"tool":39,"command":40},"pnpm","pnpm dlx skills add graysurf\u002Fagent-runtime-kit --skill execution-capsule","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fgraysurf\u002Fagent-runtime-kit\u002Fblob\u002Fmain\u002Ftests\u002Fgolden\u002Fhermes\u002Fplugins\u002Fmeta\u002Fskills\u002Fexecution-capsule\u002Fexpected\u002FSKILL.md",true,[],[45,51,56,63,70,77],{"skillNo":46,"name":47,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":48,"categories":49,"keywords":50,"publishedAt":18},"01M736XWK14DFRERMQ8RT3HS0Z","removable-skill","https:\u002F\u002Fgithub.com\u002Fgraysurf\u002Fagent-runtime-kit\u002Fblob\u002Fmain\u002Ftests\u002Fruntime-smoke\u002Ffixtures\u002Fskill-lifecycle\u002Fremove-skill\u002Ftests\u002Fgolden\u002Fcodex\u002Fplugins\u002Ffixture\u002Fskills\u002Fremovable-skill\u002FSKILL.md",[],[],{"skillNo":52,"name":47,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":53,"categories":54,"keywords":55,"publishedAt":18},"01M736XTPQEKAYESDM1REN2KAN","https:\u002F\u002Fgithub.com\u002Fgraysurf\u002Fagent-runtime-kit\u002Fblob\u002Fmain\u002Ftests\u002Fruntime-smoke\u002Ffixtures\u002Fskill-lifecycle\u002Fremove-skill\u002Ftests\u002Fgolden\u002Fclaude\u002Fplugins\u002Ffixture\u002Fskills\u002Fremovable-skill\u002FSKILL.md",[],[],{"skillNo":57,"name":58,"nameZh":6,"description":59,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":60,"categories":61,"keywords":62,"publishedAt":18},"01M736XRMWHMJB4X563G9J6QYY","removable-project-skill","Fixture project-local skill used to prove remove-project-skill dry-run coverage.","https:\u002F\u002Fgithub.com\u002Fgraysurf\u002Fagent-runtime-kit\u002Fblob\u002Fmain\u002Ftests\u002Fruntime-smoke\u002Ffixtures\u002Fskill-lifecycle\u002Fremove-project-skill\u002F.agents\u002Fskills\u002Fremovable-project-skill\u002FSKILL.md",[],[],{"skillNo":64,"name":65,"nameZh":6,"description":66,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":67,"categories":68,"keywords":69,"publishedAt":18},"01M736XPT65YZMPCBD4K3RHV4P","project-sample-skill","Fixture project-local skill used to prove create-project-skill governance coverage.","https:\u002F\u002Fgithub.com\u002Fgraysurf\u002Fagent-runtime-kit\u002Fblob\u002Fmain\u002Ftests\u002Fruntime-smoke\u002Ffixtures\u002Fskill-lifecycle\u002Fcreate-project-skill\u002F.agents\u002Fskills\u002Fproject-sample-skill\u002FSKILL.md",[],[],{"skillNo":71,"name":72,"nameZh":6,"description":73,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":74,"categories":75,"keywords":76,"publishedAt":18},"01M736XKZVYWVQA1YB0P459WND","project-local-skill","Fixture project-local skill used by runtime-kit project-local smoke checks.","https:\u002F\u002Fgithub.com\u002Fgraysurf\u002Fagent-runtime-kit\u002Fblob\u002Fmain\u002Ftests\u002Fprojects\u002Fproject-local-smoke\u002F.agents\u002Fskills\u002Fproject-local-skill\u002FSKILL.md",[],[],{"skillNo":78,"name":79,"nameZh":6,"description":80,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":81,"categories":82,"keywords":83,"publishedAt":18},"01M736XHA73KZ5G2TKWSED5P1B","topic-radar","Aggregate read-only AI and technology trend signals (Polymarket, Hacker News, GitHub, arXiv, Hugging Face, news search) into source-grounded Markdown or JSON; the lower-level data engine behind `daily-brief`.","https:\u002F\u002Fgithub.com\u002Fgraysurf\u002Fagent-runtime-kit\u002Fblob\u002Fmain\u002Ftests\u002Fgolden\u002Fhermes\u002Fplugins\u002Freporting\u002Fskills\u002Ftopic-radar\u002Fexpected\u002FSKILL.md",[],[]]