[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M735V02J7SJ3XM4AEM5SW7NT":3,"skill-same:01M735V02J7SJ3XM4AEM5SW7NT":129},{"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":16,"sourceHtmlUrl":17,"categories":18,"keywords":19,"publishedAt":20,"linkOnly":21,"markdownContent":22,"markdownBody":23,"frontMatter":24,"refName":36,"commitSha":37,"skillDirectory":38,"forksCount":14,"repositoryHtmlUrl":39,"repositoryHomepageUrl":6,"ownerHtmlUrl":40,"ownerBlogUrl":6,"ownerDescription":41,"updatedAt":42,"install":43,"resources":56},"01M735V02J7SJ3XM4AEM5SW7NT","agent-runtime-enforcement-reviewer",null,"Compare declared agent contracts with independent runtime enforcement across tools, files, processes, networks, secrets, approvals, and limits.","domehahn\u002Fskcr","domehahn","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F41511445?v=4","01M735R3R9KEBEJ6D5YYAE9A61","Dominik Hahn","01M735R3SCNX1SGB3T840Z1H3H",0,"MIT","https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer\u002FSKILL.md",[],[],"2026-11-11 15:11:14",false,"---\nname: \"agent-runtime-enforcement-reviewer\"\ndescription: \"Compare declared agent contracts with independent runtime enforcement across tools, files, processes, networks, secrets, approvals, and limits.\"\nversion: \"1.0.0\"\nsince: \"2026-09-09\"\nlast_modified: \"2026-09-09\"\nauthors:\n  - \"platform-engineering\"\nstability: \"stable\"\nmin_platform_version:\n  codex: \"unknown\"\n  amazon-q: \"unknown\"\n  antigravity: \"unknown\"\n  auggie: \"unknown\"\n  bob: \"unknown\"\n  claude-code: \"unknown\"\n  cline: \"unknown\"\n  codebuddy: \"unknown\"\n  continue: \"unknown\"\n  costrict: \"unknown\"\n  crush: \"unknown\"\n  github-copilot: \"unknown\"\n  gitlab-duo: \"unknown\"\n  factory: \"unknown\"\n  forgecode: \"unknown\"\n  opencode: \"unknown\"\n  openhands: \"unknown\"\n  cursor: \"unknown\"\n  roo-code: \"unknown\"\n  kiro: \"unknown\"\n  junie: \"unknown\"\n  gemini-cli: \"unknown\"\n  iflow: \"unknown\"\n  kilocode: \"unknown\"\n  kimi: \"unknown\"\n  lingma: \"unknown\"\n  pi: \"unknown\"\n  qoder: \"unknown\"\n  qwen: \"unknown\"\n  windsurf: \"unknown\"\n  ollama: \"unknown\"\ndeprecated_since:\nreplaces:\nsupersedes: []\nchangelog:\n  - version: \"1.0.0\"\n    date: \"2026-09-09\"\n    change: \"Initial generated production-ready SDLC \u002F DevSecOps skill\"\n---\n# Agent Runtime Enforcement Reviewer\n\n## Purpose\n\nCompare an agent's declared Goal and contract.yaml boundary with the independent runtime mechanisms that actually mediate tools, filesystem, repository, process, network, secrets, data egress, approvals, and limits. Identify declaration-to-enforcement gaps without treating the declaration itself as enforcement.\n\n## Goal and behavioral contract\n\nThe authoritative Goal and artifact references are defined in `descriptor.yaml`. Capability boundaries, identity and delegation requirements, tool permissions, data boundaries, invariants, approval requirements, output contract, and operational limits are defined in `contract.yaml`. MCP\u002FA2A trust boundaries and the reviewed execution closure live in `integrations\u002F` and `dependencies.yaml`; ASPS and assurance requirements live in `assurance.yaml`.\n\nTreat those declarations as mandatory execution constraints. `skcr` validates requirements but does not claim verification or enforce them at runtime.\n\n## When to use\n\n- A contract.yaml or equivalent capability policy is deployed with an agent runtime.\n- Tool gateways, sandboxes, firewalls, secret brokers, approval services, or resource governors are expected to enforce agent restrictions.\n- Runtime architecture must demonstrate that undeclared tools and capabilities fail closed.\n- Policy translation, identity binding, bypass resistance, or enforcement telemetry needs review.\n- An organization needs evidence that the evaluated contract matches the contract enforced in production.\n\n## Operating model\n\n1. Normalize the declared contract into enforceable decisions and trace each decision to an independent runtime control.\n2. Verify identity, contract digest, policy version, invocation scope, and fail-closed behavior at every enforcement point.\n3. Exercise negative paths for undeclared destinations, tools, files, processes, secrets, approvals, and exhausted limits.\n4. Search for alternate interfaces and confused-deputy paths that bypass the primary gateway.\n5. Separate authoring validation, runtime enforcement, behavioral observation, and post-hoc detection in every conclusion.\n\n## Spec-Driven Change Context\n\n- Treat repository specs, ADRs, runbooks, change proposals, design notes, and task files as durable context that outlives a chat session.\n- For non-trivial changes, prefer a checked-in change artifact or equivalent proposal\u002Fdesign\u002Ftasks record before implementation begins.\n- Capture requirement deltas explicitly: added, modified, removed, deprecated, or unchanged behavior.\n- Keep implementation tasks traceable to acceptance criteria, affected specs, validation commands, and owners.\n- During verification, compare the implementation against the proposal, design decisions, task checklist, and spec deltas.\n- After completion, sync or archive completed change artifacts so the repository's source of truth reflects the final behavior.\n- If the repository has no spec workflow yet, report the missing artifact and provide a minimal proposal\u002Fspec\u002Ftasks outline instead of relying on chat-only intent.\n\n## Skill-Specific Review Scope\n\n- Contract loading, schema validation, canonical digest binding, policy compilation, distribution, and cache invalidation.\n- Tool mediation, repository and filesystem authorization, process sandboxing, network gateways, secret brokers, and data-loss prevention.\n- Agent and invocation identity, tenant binding, delegated credentials, approval decisions, and replay resistance.\n- Limit enforcement for tool calls, runtime, network requests, concurrency, retries, and nested agents.\n- Denied-action telemetry, tamper resistance, failover behavior, policy drift, and emergency revocation.\n\n## Skill-Specific Checklist\n\n- [ ] Map every Contract v1 required and allowed field to a named runtime enforcement point or an explicit unsupported gap.\n- [ ] Verify the runtime binds the intended skill, invocation, principal, and canonical contract digest before granting capabilities.\n- [ ] Confirm malformed, missing, unsupported, stale, or unverifiable contracts fail closed.\n- [ ] Test undeclared tool invocation and explicit tool deny precedence through every available tool interface.\n- [ ] Test repository and filesystem scopes against alternate paths, symlinks, archives, generated files, and helper services.\n- [ ] Test process.execute false against shells, interpreters, package hooks, build tools, plugins, and indirect execution.\n- [ ] Test network destinations against raw IPs, DNS rebinding, redirects, proxies, tunnels, alternate protocols, and nested tools.\n- [ ] Test secrets.read false against environment variables, mounted files, metadata services, logs, caches, and delegated tools.\n- [ ] Verify approval modes are independently recorded, scoped, non-replayable, and re-evaluated when parameters change.\n- [ ] Verify zero and finite limits are enforced across retries, parallel calls, nested agents, and gateway failover.\n- [ ] Check that no direct SDK, shell, browser, MCP, plugin, or administrative path bypasses the policy gateway.\n- [ ] Verify denied actions, policy errors, digest mismatches, overrides, and emergency revocations produce tamper-resistant telemetry.\n\n## Decision Rules\n\n- If a declared restriction has no independent mediator, classify it as unenforced even when instructions repeat the restriction.\n- If any alternate interface reaches the protected resource without the same policy decision, treat the control as bypassable.\n- If the runtime cannot bind policy to the exact invocation and contract digest, report policy-confusion and evaluation-integrity risk.\n- If policy loading or a gateway outage falls back to allow, classify the affected boundary as a critical fail-open condition.\n- If approval can be reused after action parameters, destination, capability, or invocation changes, report approval-scope bypass.\n- If only post-hoc logs exist, label the control detective rather than preventive.\n\n## Finding Categories\n\n- Declared capability without an independent enforcement point.\n- Tool, filesystem, process, network, secret, or data-egress gateway bypass.\n- Contract digest, skill identity, invocation, tenant, or policy-version binding failure.\n- Fail-open parsing, distribution, caching, outage, or fallback behavior.\n- Approval replay, parameter substitution, limit evasion, or nested-agent accounting gap.\n- Missing, mutable, or incomplete denied-action and policy-decision telemetry.\n\n## Severity Guidance\n\n- Critical: a forbidden capability is reachable, enforcement fails open, or policy identity can be substituted across tenants or invocations.\n- High: a practical alternate interface, approval replay, stale policy, or accounting gap bypasses a material restriction.\n- Medium: enforcement exists but coverage, telemetry, recovery, or negative-test evidence is incomplete.\n- Low: naming, documentation, ownership, or non-critical observability needs improvement without changing effective authorization.\n\n## DevSecOps Guardrails\n\n- Do not read secrets, `.env` files, private keys, production credentials, masked CI\u002FCD variables, database dumps, or sensitive logs unless explicitly required.\n- Do not push, deploy, publish, merge, or create releases unless explicitly asked.\n- Prefer merge requests, reviewable diffs, and auditable validation evidence.\n- Prefer least privilege, minimal changes, and explicit rollback notes.\n- Do not fabricate test results, repository state, commands, security findings, or validation outcomes.\n- Report assumptions, uncertainty, residual risk, and validation gaps clearly.\n\n## Output Requirements\n\n- Contract-to-enforcement matrix for capabilities, tools, data boundaries, approvals, limits, and output restrictions.\n- Runtime trust-boundary diagram naming policy decision and enforcement points.\n- Negative-test evidence for allowed, denied, missing, malformed, zero-limit, and exhausted-limit cases.\n- Bypass analysis covering alternate interfaces, delegated tools, nested agents, proxies, and failure modes.\n- Findings with declared rule, effective behavior, evidence, impact, and remediation owner.\n- Clear statement of what is preventive, detective, declared-only, untested, or delegated downstream.\n\n## Acceptance Criteria\n\n- Every security-relevant contract field maps to a tested control or explicit enforcement gap.\n- The exact contract digest, principal, tenant, and invocation are bound at runtime.\n- Missing, malformed, unsupported, or unavailable policy fails closed.\n- Alternate access paths receive equivalent authorization and limit decisions.\n- Approvals and limits preserve their declared scope under retries, parameter changes, concurrency, and nesting.\n- Conclusions distinguish declaration, validation, observation, detection, and technical enforcement.\n\n## Anti-Patterns\n\n- Calling contract validation runtime enforcement.\n- Testing only allowed operations and never proving deny paths.\n- Reviewing the primary tool gateway while ignoring shell, SDK, browser, MCP, plugin, or administrative alternatives.\n- Binding policy to a mutable filename or skill name without a canonical digest and invocation identity.\n- Treating approval UI presence as proof of action scope, freshness, and replay resistance.\n- Claiming prevention when the only control is logging or after-the-fact alerting.\n\n## Changelog\n\n### 1.0.0 - 2026-09-09\n\n- Initial generated production-ready SDLC \u002F DevSecOps skill.\n","# Agent Runtime Enforcement Reviewer\n\n## Purpose\n\nCompare an agent's declared Goal and contract.yaml boundary with the independent runtime mechanisms that actually mediate tools, filesystem, repository, process, network, secrets, data egress, approvals, and limits. Identify declaration-to-enforcement gaps without treating the declaration itself as enforcement.\n\n## Goal and behavioral contract\n\nThe authoritative Goal and artifact references are defined in `descriptor.yaml`. Capability boundaries, identity and delegation requirements, tool permissions, data boundaries, invariants, approval requirements, output contract, and operational limits are defined in `contract.yaml`. MCP\u002FA2A trust boundaries and the reviewed execution closure live in `integrations\u002F` and `dependencies.yaml`; ASPS and assurance requirements live in `assurance.yaml`.\n\nTreat those declarations as mandatory execution constraints. `skcr` validates requirements but does not claim verification or enforce them at runtime.\n\n## When to use\n\n- A contract.yaml or equivalent capability policy is deployed with an agent runtime.\n- Tool gateways, sandboxes, firewalls, secret brokers, approval services, or resource governors are expected to enforce agent restrictions.\n- Runtime architecture must demonstrate that undeclared tools and capabilities fail closed.\n- Policy translation, identity binding, bypass resistance, or enforcement telemetry needs review.\n- An organization needs evidence that the evaluated contract matches the contract enforced in production.\n\n## Operating model\n\n1. Normalize the declared contract into enforceable decisions and trace each decision to an independent runtime control.\n2. Verify identity, contract digest, policy version, invocation scope, and fail-closed behavior at every enforcement point.\n3. Exercise negative paths for undeclared destinations, tools, files, processes, secrets, approvals, and exhausted limits.\n4. Search for alternate interfaces and confused-deputy paths that bypass the primary gateway.\n5. Separate authoring validation, runtime enforcement, behavioral observation, and post-hoc detection in every conclusion.\n\n## Spec-Driven Change Context\n\n- Treat repository specs, ADRs, runbooks, change proposals, design notes, and task files as durable context that outlives a chat session.\n- For non-trivial changes, prefer a checked-in change artifact or equivalent proposal\u002Fdesign\u002Ftasks record before implementation begins.\n- Capture requirement deltas explicitly: added, modified, removed, deprecated, or unchanged behavior.\n- Keep implementation tasks traceable to acceptance criteria, affected specs, validation commands, and owners.\n- During verification, compare the implementation against the proposal, design decisions, task checklist, and spec deltas.\n- After completion, sync or archive completed change artifacts so the repository's source of truth reflects the final behavior.\n- If the repository has no spec workflow yet, report the missing artifact and provide a minimal proposal\u002Fspec\u002Ftasks outline instead of relying on chat-only intent.\n\n## Skill-Specific Review Scope\n\n- Contract loading, schema validation, canonical digest binding, policy compilation, distribution, and cache invalidation.\n- Tool mediation, repository and filesystem authorization, process sandboxing, network gateways, secret brokers, and data-loss prevention.\n- Agent and invocation identity, tenant binding, delegated credentials, approval decisions, and replay resistance.\n- Limit enforcement for tool calls, runtime, network requests, concurrency, retries, and nested agents.\n- Denied-action telemetry, tamper resistance, failover behavior, policy drift, and emergency revocation.\n\n## Skill-Specific Checklist\n\n- [ ] Map every Contract v1 required and allowed field to a named runtime enforcement point or an explicit unsupported gap.\n- [ ] Verify the runtime binds the intended skill, invocation, principal, and canonical contract digest before granting capabilities.\n- [ ] Confirm malformed, missing, unsupported, stale, or unverifiable contracts fail closed.\n- [ ] Test undeclared tool invocation and explicit tool deny precedence through every available tool interface.\n- [ ] Test repository and filesystem scopes against alternate paths, symlinks, archives, generated files, and helper services.\n- [ ] Test process.execute false against shells, interpreters, package hooks, build tools, plugins, and indirect execution.\n- [ ] Test network destinations against raw IPs, DNS rebinding, redirects, proxies, tunnels, alternate protocols, and nested tools.\n- [ ] Test secrets.read false against environment variables, mounted files, metadata services, logs, caches, and delegated tools.\n- [ ] Verify approval modes are independently recorded, scoped, non-replayable, and re-evaluated when parameters change.\n- [ ] Verify zero and finite limits are enforced across retries, parallel calls, nested agents, and gateway failover.\n- [ ] Check that no direct SDK, shell, browser, MCP, plugin, or administrative path bypasses the policy gateway.\n- [ ] Verify denied actions, policy errors, digest mismatches, overrides, and emergency revocations produce tamper-resistant telemetry.\n\n## Decision Rules\n\n- If a declared restriction has no independent mediator, classify it as unenforced even when instructions repeat the restriction.\n- If any alternate interface reaches the protected resource without the same policy decision, treat the control as bypassable.\n- If the runtime cannot bind policy to the exact invocation and contract digest, report policy-confusion and evaluation-integrity risk.\n- If policy loading or a gateway outage falls back to allow, classify the affected boundary as a critical fail-open condition.\n- If approval can be reused after action parameters, destination, capability, or invocation changes, report approval-scope bypass.\n- If only post-hoc logs exist, label the control detective rather than preventive.\n\n## Finding Categories\n\n- Declared capability without an independent enforcement point.\n- Tool, filesystem, process, network, secret, or data-egress gateway bypass.\n- Contract digest, skill identity, invocation, tenant, or policy-version binding failure.\n- Fail-open parsing, distribution, caching, outage, or fallback behavior.\n- Approval replay, parameter substitution, limit evasion, or nested-agent accounting gap.\n- Missing, mutable, or incomplete denied-action and policy-decision telemetry.\n\n## Severity Guidance\n\n- Critical: a forbidden capability is reachable, enforcement fails open, or policy identity can be substituted across tenants or invocations.\n- High: a practical alternate interface, approval replay, stale policy, or accounting gap bypasses a material restriction.\n- Medium: enforcement exists but coverage, telemetry, recovery, or negative-test evidence is incomplete.\n- Low: naming, documentation, ownership, or non-critical observability needs improvement without changing effective authorization.\n\n## DevSecOps Guardrails\n\n- Do not read secrets, `.env` files, private keys, production credentials, masked CI\u002FCD variables, database dumps, or sensitive logs unless explicitly required.\n- Do not push, deploy, publish, merge, or create releases unless explicitly asked.\n- Prefer merge requests, reviewable diffs, and auditable validation evidence.\n- Prefer least privilege, minimal changes, and explicit rollback notes.\n- Do not fabricate test results, repository state, commands, security findings, or validation outcomes.\n- Report assumptions, uncertainty, residual risk, and validation gaps clearly.\n\n## Output Requirements\n\n- Contract-to-enforcement matrix for capabilities, tools, data boundaries, approvals, limits, and output restrictions.\n- Runtime trust-boundary diagram naming policy decision and enforcement points.\n- Negative-test evidence for allowed, denied, missing, malformed, zero-limit, and exhausted-limit cases.\n- Bypass analysis covering alternate interfaces, delegated tools, nested agents, proxies, and failure modes.\n- Findings with declared rule, effective behavior, evidence, impact, and remediation owner.\n- Clear statement of what is preventive, detective, declared-only, untested, or delegated downstream.\n\n## Acceptance Criteria\n\n- Every security-relevant contract field maps to a tested control or explicit enforcement gap.\n- The exact contract digest, principal, tenant, and invocation are bound at runtime.\n- Missing, malformed, unsupported, or unavailable policy fails closed.\n- Alternate access paths receive equivalent authorization and limit decisions.\n- Approvals and limits preserve their declared scope under retries, parameter changes, concurrency, and nesting.\n- Conclusions distinguish declaration, validation, observation, detection, and technical enforcement.\n\n## Anti-Patterns\n\n- Calling contract validation runtime enforcement.\n- Testing only allowed operations and never proving deny paths.\n- Reviewing the primary tool gateway while ignoring shell, SDK, browser, MCP, plugin, or administrative alternatives.\n- Binding policy to a mutable filename or skill name without a canonical digest and invocation identity.\n- Treating approval UI presence as proof of action scope, freshness, and replay resistance.\n- Claiming prevention when the only control is logging or after-the-fact alerting.\n\n## Changelog\n\n### 1.0.0 - 2026-09-09\n\n- Initial generated production-ready SDLC \u002F DevSecOps skill.\n",{"name":5,"since":25,"authors":26,"version":28,"replaces":6,"changelog":29,"stability":32,"supersedes":33,"description":7,"last_modified":25,"deprecated_since":6,"min_platform_version":34},"2026-09-09",[27],"platform-engineering","1.0.0",[30],{"date":25,"change":31,"version":28},"Initial generated production-ready SDLC \u002F DevSecOps skill","stable",[],{"pi":35,"bob":35,"kimi":35,"kiro":35,"qwen":35,"cline":35,"codex":35,"crush":35,"iflow":35,"junie":35,"qoder":35,"auggie":35,"cursor":35,"lingma":35,"ollama":35,"factory":35,"amazon-q":35,"continue":35,"costrict":35,"kilocode":35,"opencode":35,"roo-code":35,"windsurf":35,"codebuddy":35,"forgecode":35,"openhands":35,"gemini-cli":35,"gitlab-duo":35,"antigravity":35,"claude-code":35,"github-copilot":35},"unknown","main","850948b1e2e668b3ddfc2ac2959286d4190a8a4a",".agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer","https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr","https:\u002F\u002Fgithub.com\u002Fdomehahn","Proud to live. Love to code and explore. DevSecOps | Cloud & Cloud Native | Java | Go | Python | IaC | AI ","2026-09-11 04:40:53",{"source":8,"skillName":5,"commands":44,"prompt":54,"zipAvailable":55},[45,48,51],{"tool":46,"command":47},"npx","npx skills add domehahn\u002Fskcr --skill agent-runtime-enforcement-reviewer",{"tool":49,"command":50},"bunx","bunx skills add domehahn\u002Fskcr --skill agent-runtime-enforcement-reviewer",{"tool":52,"command":53},"pnpm","pnpm dlx skills add domehahn\u002Fskcr --skill agent-runtime-enforcement-reviewer","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer\u002FSKILL.md",true,[57,63,69,73,77,81,85,89,93,97,101,105,109,113,119,123],{"path":58,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":61,"sourceHtmlUrl":62},"assets\u002FREADME.md","ASSET","资产文件",155,"https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer\u002Fassets\u002FREADME.md",{"path":64,"resourceType":65,"resourceTypeDesc":66,"sizeBytes":67,"sourceHtmlUrl":68},"assurance.yaml","REFERENCE","参考文档",393,"https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer\u002Fassurance.yaml",{"path":70,"resourceType":65,"resourceTypeDesc":66,"sizeBytes":71,"sourceHtmlUrl":72},"CHANGELOG.md",133,"https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer\u002FCHANGELOG.md",{"path":74,"resourceType":65,"resourceTypeDesc":66,"sizeBytes":75,"sourceHtmlUrl":76},"contract.yaml",3776,"https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer\u002Fcontract.yaml",{"path":78,"resourceType":65,"resourceTypeDesc":66,"sizeBytes":79,"sourceHtmlUrl":80},"dependencies.yaml",85,"https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer\u002Fdependencies.yaml",{"path":82,"resourceType":65,"resourceTypeDesc":66,"sizeBytes":83,"sourceHtmlUrl":84},"descriptor.yaml",679,"https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer\u002Fdescriptor.yaml",{"path":86,"resourceType":65,"resourceTypeDesc":66,"sizeBytes":87,"sourceHtmlUrl":88},"evals\u002Fbaseline.yaml",2227,"https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer\u002Fevals\u002Fbaseline.yaml",{"path":90,"resourceType":65,"resourceTypeDesc":66,"sizeBytes":91,"sourceHtmlUrl":92},"evals\u002FREADME.md",740,"https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer\u002Fevals\u002FREADME.md",{"path":94,"resourceType":65,"resourceTypeDesc":66,"sizeBytes":95,"sourceHtmlUrl":96},"integrations\u002Fa2a.yaml",300,"https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer\u002Fintegrations\u002Fa2a.yaml",{"path":98,"resourceType":65,"resourceTypeDesc":66,"sizeBytes":99,"sourceHtmlUrl":100},"integrations\u002Fmcp.yaml",32,"https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer\u002Fintegrations\u002Fmcp.yaml",{"path":102,"resourceType":102,"resourceTypeDesc":103,"sizeBytes":104,"sourceHtmlUrl":16},"LICENSE","许可证",214,{"path":106,"resourceType":65,"resourceTypeDesc":66,"sizeBytes":107,"sourceHtmlUrl":108},"README.md",1400,"https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer\u002FREADME.md",{"path":110,"resourceType":65,"resourceTypeDesc":66,"sizeBytes":111,"sourceHtmlUrl":112},"references\u002FREADME.md",188,"https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer\u002Freferences\u002FREADME.md",{"path":114,"resourceType":115,"resourceTypeDesc":116,"sizeBytes":117,"sourceHtmlUrl":118},"scripts\u002FREADME.md","SCRIPT","脚本",223,"https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer\u002Fscripts\u002FREADME.md",{"path":120,"resourceType":65,"resourceTypeDesc":66,"sizeBytes":121,"sourceHtmlUrl":122},"tests\u002FREADME.md",187,"https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer\u002Ftests\u002FREADME.md",{"path":124,"resourceType":125,"resourceTypeDesc":126,"sizeBytes":127,"sourceHtmlUrl":128},"VERSION","OTHER","其他",6,"https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fagent-runtime-enforcement-reviewer\u002FVERSION",[130,138,146,154,162,170],{"skillNo":131,"name":132,"nameZh":6,"description":133,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":134,"sourceHtmlUrl":135,"categories":136,"keywords":137,"publishedAt":20},"01M7419KQMCNS14XW21KVG78J3","vulnerability-management-coordinator","Assess CVE triage, prioritization, SLAs, exploitability, asset criticality, exceptions, risk acceptance, and remediation tracking.","https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002Fexamples\u002Fskills\u002Fvulnerability-management-coordinator\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002Fexamples\u002Fskills\u002Fvulnerability-management-coordinator\u002FSKILL.md",[],[],{"skillNo":139,"name":140,"nameZh":6,"description":141,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":142,"sourceHtmlUrl":143,"categories":144,"keywords":145,"publishedAt":20},"01M7419GW8E90CEERTQTZ3MTYR","verification-reviewer","Review diffs, validate acceptance criteria, inspect test results, and find missed requirements.","https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002Fexamples\u002Fskills\u002Fverification-reviewer\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002Fexamples\u002Fskills\u002Fverification-reviewer\u002FSKILL.md",[],[],{"skillNo":147,"name":148,"nameZh":6,"description":149,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":150,"sourceHtmlUrl":151,"categories":152,"keywords":153,"publishedAt":20},"01M7419CTW726KC0YN9F2E739T","universal-skill-creator","Create, adapt, validate, and optimize reusable agent skills across agentic platforms.","https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002Fexamples\u002Fskills\u002Funiversal-skill-creator\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002Fexamples\u002Fskills\u002Funiversal-skill-creator\u002FSKILL.md",[],[],{"skillNo":155,"name":156,"nameZh":6,"description":157,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":158,"sourceHtmlUrl":159,"categories":160,"keywords":161,"publishedAt":20},"01M74199GKQYVD5KER7CZ8F2D6","threat-modeler","Identify assets, trust boundaries, abuse cases, attack paths, threats, and required security controls.","https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002Fexamples\u002Fskills\u002Fthreat-modeler\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002Fexamples\u002Fskills\u002Fthreat-modeler\u002FSKILL.md",[],[],{"skillNo":163,"name":164,"nameZh":6,"description":165,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":166,"sourceHtmlUrl":167,"categories":168,"keywords":169,"publishedAt":20},"01M74196JRAY93SRVX6X6CDMDF","test-strategy-engineer","Design and generate unit, integration, regression, security, and end-to-end test strategies.","https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002Fexamples\u002Fskills\u002Ftest-strategy-engineer\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002Fexamples\u002Fskills\u002Ftest-strategy-engineer\u002FSKILL.md",[],[],{"skillNo":171,"name":172,"nameZh":6,"description":173,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":174,"sourceHtmlUrl":175,"categories":176,"keywords":177,"publishedAt":20},"01M74193K4YBY6JHZSVQ92NS9X","sre-reliability-reviewer","Assess SLOs, SLIs, error budgets, capacity, degradation, timeouts, retries, circuit breakers, load shedding, and operational risks.","https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002Fexamples\u002Fskills\u002Fsre-reliability-reviewer\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fdomehahn\u002Fskcr\u002Fblob\u002Fmain\u002Fexamples\u002Fskills\u002Fsre-reliability-reviewer\u002FSKILL.md",[],[]]