[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7BXWSWDSGFR3VWS2B347B8V":3,"skill-same:01M7BXWSWDSGFR3VWS2B347B8V":74},{"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":25,"commitSha":26,"skillDirectory":27,"forksCount":28,"repositoryHtmlUrl":29,"repositoryHomepageUrl":6,"ownerHtmlUrl":30,"ownerBlogUrl":31,"ownerDescription":32,"updatedAt":33,"install":34,"resources":47},"01M7BXWSWDSGFR3VWS2B347B8V","featureops",null,"Manage feature flags with Unleash — evaluate whether a code change needs a flag, create flags with consistent naming, wrap code behind flags, plan progressive rollouts, and clean up stale flags. Use when working with feature flags, feature toggles, rollouts, releases, experiments, kill-switches, canary deploys, or the Unleash MCP server. Ships as a customizable FeatureOps template — set your team's policy in an override file (`.unleash\u002Ffeatureops.md`); no need to edit the skill.","CognitionAI\u002Fdevin-marketplace","CognitionAI","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F163493330?v=4","01M7BVSRD98NC4FG8T3EZS6V5M","Cognition","01M7BVSRDX111C2XPWJ0E6MY26",29,"Apache-2.0","https:\u002F\u002Fgithub.com\u002FCognitionAI\u002Fdevin-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Funleash\u002Fskills\u002Ffeatureops\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: featureops\ndescription: Manage feature flags with Unleash — evaluate whether a code change needs a flag, create flags with consistent naming, wrap code behind flags, plan progressive rollouts, and clean up stale flags. Use when working with feature flags, feature toggles, rollouts, releases, experiments, kill-switches, canary deploys, or the Unleash MCP server. Ships as a customizable FeatureOps template — set your team's policy in an override file (`.unleash\u002Ffeatureops.md`); no need to edit the skill.\nlicense: Apache-2.0\nallowed-tools: mcp__unleash__evaluate_change mcp__unleash__detect_flag mcp__unleash__get_flag_state mcp__unleash__list_projects mcp__unleash__list_flags mcp__plugin_unleash_mcp__evaluate_change mcp__plugin_unleash_mcp__detect_flag mcp__plugin_unleash_mcp__get_flag_state mcp__plugin_unleash_mcp__list_projects mcp__plugin_unleash_mcp__list_flags\n---\n\n# Unleash · FeatureOps Template\n\nThis skill brings [FeatureOps](https:\u002F\u002Ffeatureops.io) — the discipline of controlling software behavior at runtime — into your coding agent. It drives the [Unleash MCP server](https:\u002F\u002Fgithub.com\u002FUnleash\u002Funleash-mcp) to evaluate changes for risk, create flags with consistent naming, generate SDK guard code, manage rollouts, and clean up stale flags.\n\nThe defaults work out of the box, but feature-management conventions vary across teams and industries. This skill ships as a **customizable template**: every tunable value lives in one policy file, which you override per project *without editing the skill* — see [Active policy](#active-policy) and [Customize for your team](#customize-for-your-team).\n\n## Setup check\n\nThis skill needs the Unleash MCP server connected. When installed as a Claude Code plugin, the server is registered automatically from the bundled `.mcp.json`; you only need to set three environment variables in the shell that launches your agent. With a standalone install on any other agent, connect the Unleash MCP server yourself first (see `examples\u002FREADME.md` in the repository, https:\u002F\u002Fgithub.com\u002FUnleash\u002Funleash-claude-skills, including the remote MCP server option):\n\n```bash\nexport UNLEASH_BASE_URL=https:\u002F\u002Fyour-instance.getunleash.io\u002Fapi\nexport UNLEASH_PAT=your-personal-access-token\nexport UNLEASH_DEFAULT_PROJECT=default   # optional; defaults to \"default\"\n```\n\nIf the Unleash tools are not available, tell the user to complete setup (see the README at https:\u002F\u002Fgithub.com\u002FUnleash\u002Funleash-claude-skills for full install, transport, and credential instructions) and stop rather than guessing flag state.\n\n> **Tool naming note (Claude Code):** on Claude Code the read-only tools are pre-approved via `allowed-tools` under two names each: `mcp__unleash__*` (standalone-skill install, where you register the server under the key `unleash`) and `mcp__plugin_unleash_mcp__*` (plugin install, where Claude Code namespaces the bundled server — keyed `mcp` in `.mcp.json` — as `mcp__plugin_\u003Cplugin>_\u003Cserver>__\u003Ctool>` → `mcp__plugin_unleash_mcp__\u003Ctool>`, verified on a real install). Pre-approval only applies while this skill is active; a read-only tool invoked before the skill has loaded will still prompt once. Write operations (`create_flag`, `set_flag_rollout`, `toggle_flag_environment`, `remove_flag_strategy`) are **never** pre-approved — always let the user confirm them. `wrap_change` and `cleanup_flag` don't mutate server state (they return guard code and removal instructions), but they are deliberately not pre-approved either: each immediately precedes edits to production code paths, so the confirmation prompt stays. On other agents, `allowed-tools` doesn't apply — they use their own permission model — but the read-only vs. write distinction above still holds: never let a write tool run without user confirmation.\n\n## Active policy\n\nEvery tunable value — naming convention, flag types and lifetimes, rollout milestones and holds, halt thresholds, staleness windows, and the high-risk domains — comes from the **active policy**. The reference files describe *how* to apply these values and never restate the numbers, so always resolve them from here.\n\n**Precedence, highest first: project override → user override → shipped defaults.** For each value, use the highest layer that sets it; anything unset falls through to the defaults. The project override and the defaults are inlined below at load time. If a user override exists (`~\u002F.unleash\u002Ffeatureops.md`), read it with your file tools and slot it *between* them.\n\n**Project override** — `${CLAUDE_PROJECT_DIR}\u002F.unleash\u002Ffeatureops.md` (repo-specific; wins):\n!`cat \"${CLAUDE_PROJECT_DIR}\u002F.unleash\u002Ffeatureops.md\" 2>\u002Fdev\u002Fnull || echo \"(no project override — using user override and\u002For the defaults below)\"`\n\n**Shipped defaults** — `references\u002Fpolicy-defaults.md`:\n!`cat \"${CLAUDE_SKILL_DIR}\u002Freferences\u002Fpolicy-defaults.md\" 2>\u002Fdev\u002Fnull || echo \"(defaults file not found; load references\u002Fpolicy-defaults.md manually)\"`\n\n> **If the two blocks above are empty or show a literal `` !`…` `` command**, your agent doesn't run inline injection (this is a Claude Code feature). Read the files yourself with your file tools and apply the same precedence: the project override `.unleash\u002Ffeatureops.md`, the user override `~\u002F.unleash\u002Ffeatureops.md`, and this skill's `references\u002Fpolicy-defaults.md`.\n\n## When to load the reference files\n\nLoad the reference file that matches the task. For tasks that span workflows (e.g. \"create a flag and plan its rollout\"), load both.\n\n| User intent | Load |\n|---|---|\n| Setting up feature flags, deciding when to flag a change, naming a new flag | `references\u002Ffeature-flag-conventions.md` |\n| Working in a high-risk directory (payments, auth, data migrations, external integrations) | `references\u002Fhigh-risk-domains.md` |\n| Planning a gradual rollout, configuring percentages, setting halt conditions | `references\u002Frollout-guidance.md` |\n| Removing a flag, cleaning up after a rollout, auditing stale flags | `references\u002Fcleanup-cadence.md` |\n\n## Core workflow\n\nFor a change that might need a flag, follow this sequence. The reference files hold the detailed rules for each step.\n\n1. **Evaluate risk** — call `evaluate_change` with a description of the change before implementing anything in the active policy's flag-required categories (see [Active policy](#active-policy)). See `references\u002Ffeature-flag-conventions.md` and `references\u002Fhigh-risk-domains.md`.\n2. **Prefer reuse** — call `detect_flag` before creating anything. Reuse a matching flag rather than adding to flag sprawl.\n3. **Create** — if no match exists, call `create_flag` with a name following the active policy's naming convention and the right flag type. See `references\u002Ffeature-flag-conventions.md`.\n4. **Wrap** — call `wrap_change` for SDK-appropriate guard code. Apply it verbatim; keep the fallback branch realistic.\n5. **Roll out** — use `set_flag_rollout` and `toggle_flag_environment` to progress through milestones. See `references\u002Frollout-guidance.md`.\n6. **Clean up** — flags are temporary by design. When a rollout stabilizes or an experiment concludes, remove the flag. See `references\u002Fcleanup-cadence.md`.\n\nDo not auto-execute write operations. Present the plan, get confirmation, then act — especially for flag creation, rollout changes, and cleanup, which touch production code paths.\n\n## Migrations are a different workflow\n\nWhen the task is migrating existing flags from another system (e.g. GitLab feature flags) into Unleash, defer to the `migrate-gitlab-flags` skill (`\u002Funleash:migrate-gitlab-flags`) and do **not** apply this skill's conventions to the migrated flags — migration parity rules win: flag names carry over unchanged (never renamed to `{domain}-{feature}-{variant}`), state carries over exactly (no `evaluate_change` gate, no default rollout milestones), and strategy parameters such as `groupId` are preserved verbatim. This skill's conventions govern new flags, and they resume governing migrated flags once the migration completes.\n\n## Tool reference\n\nThe Unleash MCP server exposes the following tools.\n\n| Tool | Description | When to use |\n|---|---|---|\n| `evaluate_change` | Analyzes a code change and determines whether it should be behind a feature flag. | Before implementing risky changes |\n| `detect_flag` | Searches for existing flags that match a description to prevent duplicates. | Before creating new flags |\n| `create_flag` | Creates a new feature flag with proper naming, typing, and metadata. | When no suitable flag exists |\n| `wrap_change` | Generates framework-specific code to guard a feature behind a flag. | After creating a flag |\n| `list_projects` | Lists Unleash projects available to the configured token, with optional pagination. | Discovering available projects |\n| `list_flags` | Lists feature flags in a project (active by default; set `archived=true` for archived flags). | Auditing flag inventory; discovering existing flags before creating new ones |\n| `get_flag_state` | Returns the current state, strategies, and metadata for a flag. | Debugging, status checks |\n| `set_flag_rollout` | Configures rollout percentages and activation strategies. | Gradual releases |\n| `toggle_flag_environment` | Enables or disables a flag in a specific environment. | Testing, staged rollouts |\n| `remove_flag_strategy` | Deletes a rollout strategy from a flag. | Simplifying flag configuration |\n| `cleanup_flag` | Returns file locations and instructions for removing a flag after rollout. | After full rollout |\n\nThe read-only tools — `evaluate_change`, `detect_flag`, `get_flag_state`, `list_projects`, `list_flags` — are pre-approved via `allowed-tools`. Every other tool prompts for approval — the write operations because they mutate server state, and `wrap_change`\u002F`cleanup_flag` because their output immediately precedes code edits. Keep it that way unless the user has a specific reason to lower the bar. Use `list_projects` and `list_flags` for inventory discovery before creating new flags; `list_flags` (active + archived) is also the basis of the flag audit workflow in `references\u002Fcleanup-cadence.md`.\n\n## Customize for your team\n\nDon't edit this skill's files to customize it — in a plugin install they live in a managed cache that a plugin update overwrites. Instead, put your team's policy in an **override file**; the skill reads it automatically (see [Active policy](#active-policy)). `references\u002Fpolicy-defaults.md` lists every tunable value and doubles as the template.\n\n- **Project (recommended):** create `.\u002F.unleash\u002Ffeatureops.md` in the repo and set only the values you want to change — especially your real high-risk domains. It's committed with the repo and wins over everything.\n- **User (optional):** `~\u002F.unleash\u002Ffeatureops.md` for personal cross-repo defaults.\n\nCopy the sections you want to change out of `references\u002Fpolicy-defaults.md`, or just ask the agent: *\"set up a FeatureOps policy override for this project.\"* For company-wide standardization, the heavier alternative is to fork this repo, bake your policy into `policy-defaults.md`, and publish it as your own internal plugin — see the skill README.\n\n---\n\n## Metadata\n\n- **License:** Apache-2.0\n- **Privacy Policy:** https:\u002F\u002Fwww.getunleash.io\u002Fprivacy-policy\n- **Support:** support@getunleash.io · GitHub Issues at https:\u002F\u002Fgithub.com\u002FUnleash\u002Funleash-claude-skills\u002Fissues\n- **Source:** https:\u002F\u002Fgithub.com\u002FUnleash\u002Funleash-claude-skills\n\n### MCP server attribution\n\nThis skill configures the **Unleash MCP server**:\n\n- Package: [`@unleash\u002Fmcp`](https:\u002F\u002Fwww.npmjs.com\u002Fpackage\u002F@unleash\u002Fmcp) on npm\n- Source: https:\u002F\u002Fgithub.com\u002FUnleash\u002Funleash-mcp\n- License: Apache-2.0\n- Maintainer: Unleash (Bricks Software AS)\n- Transport: local stdio via `npx`, or remote Streamable HTTP at `https:\u002F\u002F\u003Cyour-instance>\u002Fapi\u002Fadmin\u002Fmcp`\n","# Unleash · FeatureOps Template\n\nThis skill brings [FeatureOps](https:\u002F\u002Ffeatureops.io) — the discipline of controlling software behavior at runtime — into your coding agent. It drives the [Unleash MCP server](https:\u002F\u002Fgithub.com\u002FUnleash\u002Funleash-mcp) to evaluate changes for risk, create flags with consistent naming, generate SDK guard code, manage rollouts, and clean up stale flags.\n\nThe defaults work out of the box, but feature-management conventions vary across teams and industries. This skill ships as a **customizable template**: every tunable value lives in one policy file, which you override per project *without editing the skill* — see [Active policy](#active-policy) and [Customize for your team](#customize-for-your-team).\n\n## Setup check\n\nThis skill needs the Unleash MCP server connected. When installed as a Claude Code plugin, the server is registered automatically from the bundled `.mcp.json`; you only need to set three environment variables in the shell that launches your agent. With a standalone install on any other agent, connect the Unleash MCP server yourself first (see `examples\u002FREADME.md` in the repository, https:\u002F\u002Fgithub.com\u002FUnleash\u002Funleash-claude-skills, including the remote MCP server option):\n\n```bash\nexport UNLEASH_BASE_URL=https:\u002F\u002Fyour-instance.getunleash.io\u002Fapi\nexport UNLEASH_PAT=your-personal-access-token\nexport UNLEASH_DEFAULT_PROJECT=default   # optional; defaults to \"default\"\n```\n\nIf the Unleash tools are not available, tell the user to complete setup (see the README at https:\u002F\u002Fgithub.com\u002FUnleash\u002Funleash-claude-skills for full install, transport, and credential instructions) and stop rather than guessing flag state.\n\n> **Tool naming note (Claude Code):** on Claude Code the read-only tools are pre-approved via `allowed-tools` under two names each: `mcp__unleash__*` (standalone-skill install, where you register the server under the key `unleash`) and `mcp__plugin_unleash_mcp__*` (plugin install, where Claude Code namespaces the bundled server — keyed `mcp` in `.mcp.json` — as `mcp__plugin_\u003Cplugin>_\u003Cserver>__\u003Ctool>` → `mcp__plugin_unleash_mcp__\u003Ctool>`, verified on a real install). Pre-approval only applies while this skill is active; a read-only tool invoked before the skill has loaded will still prompt once. Write operations (`create_flag`, `set_flag_rollout`, `toggle_flag_environment`, `remove_flag_strategy`) are **never** pre-approved — always let the user confirm them. `wrap_change` and `cleanup_flag` don't mutate server state (they return guard code and removal instructions), but they are deliberately not pre-approved either: each immediately precedes edits to production code paths, so the confirmation prompt stays. On other agents, `allowed-tools` doesn't apply — they use their own permission model — but the read-only vs. write distinction above still holds: never let a write tool run without user confirmation.\n\n## Active policy\n\nEvery tunable value — naming convention, flag types and lifetimes, rollout milestones and holds, halt thresholds, staleness windows, and the high-risk domains — comes from the **active policy**. The reference files describe *how* to apply these values and never restate the numbers, so always resolve them from here.\n\n**Precedence, highest first: project override → user override → shipped defaults.** For each value, use the highest layer that sets it; anything unset falls through to the defaults. The project override and the defaults are inlined below at load time. If a user override exists (`~\u002F.unleash\u002Ffeatureops.md`), read it with your file tools and slot it *between* them.\n\n**Project override** — `${CLAUDE_PROJECT_DIR}\u002F.unleash\u002Ffeatureops.md` (repo-specific; wins):\n!`cat \"${CLAUDE_PROJECT_DIR}\u002F.unleash\u002Ffeatureops.md\" 2>\u002Fdev\u002Fnull || echo \"(no project override — using user override and\u002For the defaults below)\"`\n\n**Shipped defaults** — `references\u002Fpolicy-defaults.md`:\n!`cat \"${CLAUDE_SKILL_DIR}\u002Freferences\u002Fpolicy-defaults.md\" 2>\u002Fdev\u002Fnull || echo \"(defaults file not found; load references\u002Fpolicy-defaults.md manually)\"`\n\n> **If the two blocks above are empty or show a literal `` !`…` `` command**, your agent doesn't run inline injection (this is a Claude Code feature). Read the files yourself with your file tools and apply the same precedence: the project override `.unleash\u002Ffeatureops.md`, the user override `~\u002F.unleash\u002Ffeatureops.md`, and this skill's `references\u002Fpolicy-defaults.md`.\n\n## When to load the reference files\n\nLoad the reference file that matches the task. For tasks that span workflows (e.g. \"create a flag and plan its rollout\"), load both.\n\n| User intent | Load |\n|---|---|\n| Setting up feature flags, deciding when to flag a change, naming a new flag | `references\u002Ffeature-flag-conventions.md` |\n| Working in a high-risk directory (payments, auth, data migrations, external integrations) | `references\u002Fhigh-risk-domains.md` |\n| Planning a gradual rollout, configuring percentages, setting halt conditions | `references\u002Frollout-guidance.md` |\n| Removing a flag, cleaning up after a rollout, auditing stale flags | `references\u002Fcleanup-cadence.md` |\n\n## Core workflow\n\nFor a change that might need a flag, follow this sequence. The reference files hold the detailed rules for each step.\n\n1. **Evaluate risk** — call `evaluate_change` with a description of the change before implementing anything in the active policy's flag-required categories (see [Active policy](#active-policy)). See `references\u002Ffeature-flag-conventions.md` and `references\u002Fhigh-risk-domains.md`.\n2. **Prefer reuse** — call `detect_flag` before creating anything. Reuse a matching flag rather than adding to flag sprawl.\n3. **Create** — if no match exists, call `create_flag` with a name following the active policy's naming convention and the right flag type. See `references\u002Ffeature-flag-conventions.md`.\n4. **Wrap** — call `wrap_change` for SDK-appropriate guard code. Apply it verbatim; keep the fallback branch realistic.\n5. **Roll out** — use `set_flag_rollout` and `toggle_flag_environment` to progress through milestones. See `references\u002Frollout-guidance.md`.\n6. **Clean up** — flags are temporary by design. When a rollout stabilizes or an experiment concludes, remove the flag. See `references\u002Fcleanup-cadence.md`.\n\nDo not auto-execute write operations. Present the plan, get confirmation, then act — especially for flag creation, rollout changes, and cleanup, which touch production code paths.\n\n## Migrations are a different workflow\n\nWhen the task is migrating existing flags from another system (e.g. GitLab feature flags) into Unleash, defer to the `migrate-gitlab-flags` skill (`\u002Funleash:migrate-gitlab-flags`) and do **not** apply this skill's conventions to the migrated flags — migration parity rules win: flag names carry over unchanged (never renamed to `{domain}-{feature}-{variant}`), state carries over exactly (no `evaluate_change` gate, no default rollout milestones), and strategy parameters such as `groupId` are preserved verbatim. This skill's conventions govern new flags, and they resume governing migrated flags once the migration completes.\n\n## Tool reference\n\nThe Unleash MCP server exposes the following tools.\n\n| Tool | Description | When to use |\n|---|---|---|\n| `evaluate_change` | Analyzes a code change and determines whether it should be behind a feature flag. | Before implementing risky changes |\n| `detect_flag` | Searches for existing flags that match a description to prevent duplicates. | Before creating new flags |\n| `create_flag` | Creates a new feature flag with proper naming, typing, and metadata. | When no suitable flag exists |\n| `wrap_change` | Generates framework-specific code to guard a feature behind a flag. | After creating a flag |\n| `list_projects` | Lists Unleash projects available to the configured token, with optional pagination. | Discovering available projects |\n| `list_flags` | Lists feature flags in a project (active by default; set `archived=true` for archived flags). | Auditing flag inventory; discovering existing flags before creating new ones |\n| `get_flag_state` | Returns the current state, strategies, and metadata for a flag. | Debugging, status checks |\n| `set_flag_rollout` | Configures rollout percentages and activation strategies. | Gradual releases |\n| `toggle_flag_environment` | Enables or disables a flag in a specific environment. | Testing, staged rollouts |\n| `remove_flag_strategy` | Deletes a rollout strategy from a flag. | Simplifying flag configuration |\n| `cleanup_flag` | Returns file locations and instructions for removing a flag after rollout. | After full rollout |\n\nThe read-only tools — `evaluate_change`, `detect_flag`, `get_flag_state`, `list_projects`, `list_flags` — are pre-approved via `allowed-tools`. Every other tool prompts for approval — the write operations because they mutate server state, and `wrap_change`\u002F`cleanup_flag` because their output immediately precedes code edits. Keep it that way unless the user has a specific reason to lower the bar. Use `list_projects` and `list_flags` for inventory discovery before creating new flags; `list_flags` (active + archived) is also the basis of the flag audit workflow in `references\u002Fcleanup-cadence.md`.\n\n## Customize for your team\n\nDon't edit this skill's files to customize it — in a plugin install they live in a managed cache that a plugin update overwrites. Instead, put your team's policy in an **override file**; the skill reads it automatically (see [Active policy](#active-policy)). `references\u002Fpolicy-defaults.md` lists every tunable value and doubles as the template.\n\n- **Project (recommended):** create `.\u002F.unleash\u002Ffeatureops.md` in the repo and set only the values you want to change — especially your real high-risk domains. It's committed with the repo and wins over everything.\n- **User (optional):** `~\u002F.unleash\u002Ffeatureops.md` for personal cross-repo defaults.\n\nCopy the sections you want to change out of `references\u002Fpolicy-defaults.md`, or just ask the agent: *\"set up a FeatureOps policy override for this project.\"* For company-wide standardization, the heavier alternative is to fork this repo, bake your policy into `policy-defaults.md`, and publish it as your own internal plugin — see the skill README.\n\n---\n\n## Metadata\n\n- **License:** Apache-2.0\n- **Privacy Policy:** https:\u002F\u002Fwww.getunleash.io\u002Fprivacy-policy\n- **Support:** support@getunleash.io · GitHub Issues at https:\u002F\u002Fgithub.com\u002FUnleash\u002Funleash-claude-skills\u002Fissues\n- **Source:** https:\u002F\u002Fgithub.com\u002FUnleash\u002Funleash-claude-skills\n\n### MCP server attribution\n\nThis skill configures the **Unleash MCP server**:\n\n- Package: [`@unleash\u002Fmcp`](https:\u002F\u002Fwww.npmjs.com\u002Fpackage\u002F@unleash\u002Fmcp) on npm\n- Source: https:\u002F\u002Fgithub.com\u002FUnleash\u002Funleash-mcp\n- License: Apache-2.0\n- Maintainer: Unleash (Bricks Software AS)\n- Transport: local stdio via `npx`, or remote Streamable HTTP at `https:\u002F\u002F\u003Cyour-instance>\u002Fapi\u002Fadmin\u002Fmcp`\n",{"name":5,"license":15,"description":7,"allowed-tools":24},"mcp__unleash__evaluate_change mcp__unleash__detect_flag mcp__unleash__get_flag_state mcp__unleash__list_projects mcp__unleash__list_flags mcp__plugin_unleash_mcp__evaluate_change mcp__plugin_unleash_mcp__detect_flag mcp__plugin_unleash_mcp__get_flag_state mcp__plugin_unleash_mcp__list_projects mcp__plugin_unleash_mcp__list_flags","main","9836d94e88404aabb682086405c2aa9cba7657ac","plugins\u002Funleash\u002Fskills\u002Ffeatureops",156,"https:\u002F\u002Fgithub.com\u002FCognitionAI\u002Fdevin-marketplace","https:\u002F\u002Fgithub.com\u002FCognitionAI","https:\u002F\u002Fcognition.ai\u002F","Makers of Devin, the AI software engineer.","2026-11-13 07:08:23",{"source":8,"skillName":5,"commands":35,"prompt":45,"zipAvailable":46},[36,39,42],{"tool":37,"command":38},"npx","npx skills add CognitionAI\u002Fdevin-marketplace --skill featureops",{"tool":40,"command":41},"bunx","bunx skills add CognitionAI\u002Fdevin-marketplace --skill featureops",{"tool":43,"command":44},"pnpm","pnpm dlx skills add CognitionAI\u002Fdevin-marketplace --skill featureops","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FCognitionAI\u002Fdevin-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Funleash\u002Fskills\u002Ffeatureops\u002FSKILL.md",true,[48,54,58,62,66,70],{"path":49,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":52,"sourceHtmlUrl":53},"README.md","REFERENCE","参考文档",4338,"https:\u002F\u002Fgithub.com\u002FCognitionAI\u002Fdevin-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Funleash\u002Fskills\u002Ffeatureops\u002FREADME.md",{"path":55,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":56,"sourceHtmlUrl":57},"references\u002Fcleanup-cadence.md",8472,"https:\u002F\u002Fgithub.com\u002FCognitionAI\u002Fdevin-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Funleash\u002Fskills\u002Ffeatureops\u002Freferences\u002Fcleanup-cadence.md",{"path":59,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":60,"sourceHtmlUrl":61},"references\u002Ffeature-flag-conventions.md",2871,"https:\u002F\u002Fgithub.com\u002FCognitionAI\u002Fdevin-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Funleash\u002Fskills\u002Ffeatureops\u002Freferences\u002Ffeature-flag-conventions.md",{"path":63,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":64,"sourceHtmlUrl":65},"references\u002Fhigh-risk-domains.md",2284,"https:\u002F\u002Fgithub.com\u002FCognitionAI\u002Fdevin-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Funleash\u002Fskills\u002Ffeatureops\u002Freferences\u002Fhigh-risk-domains.md",{"path":67,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":68,"sourceHtmlUrl":69},"references\u002Fpolicy-defaults.md",6761,"https:\u002F\u002Fgithub.com\u002FCognitionAI\u002Fdevin-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Funleash\u002Fskills\u002Ffeatureops\u002Freferences\u002Fpolicy-defaults.md",{"path":71,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":72,"sourceHtmlUrl":73},"references\u002Frollout-guidance.md",3239,"https:\u002F\u002Fgithub.com\u002FCognitionAI\u002Fdevin-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Funleash\u002Fskills\u002Ffeatureops\u002Freferences\u002Frollout-guidance.md",[75,82,89,95,102,109],{"skillNo":76,"name":77,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":78,"licenseFileUrl":6,"sourceHtmlUrl":79,"categories":80,"keywords":81,"publishedAt":19},"01M7BYC4AJGM98P2G9JH7QS24Z","business-data","MIT","https:\u002F\u002Fgithub.com\u002FCognitionAI\u002Fdevin-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Fwindsor-ai\u002Fskills\u002Fbusiness-data\u002FSKILL.md",[],[],{"skillNo":83,"name":84,"nameZh":6,"description":85,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":78,"licenseFileUrl":6,"sourceHtmlUrl":86,"categories":87,"keywords":88,"publishedAt":19},"01M7BYBZ0CD4CBCPZHRZ551BNK","workflow","Creates durable, resumable workflows using Vercel's Workflow SDK. Use when building workflows that need to survive restarts, pause for external events, retry on failure, or coordinate multi-step operations over time. Triggers on mentions of \"workflow\", \"durable functions\", \"resumable\", \"workflow sdk\", \"queue\", \"event\", \"push\", \"subscribe\", or step-based orchestration.","https:\u002F\u002Fgithub.com\u002FCognitionAI\u002Fdevin-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Fvercel\u002Fskills\u002Fworkflow\u002Fupstream\u002FSKILL.md",[],[],{"skillNo":90,"name":84,"nameZh":6,"description":91,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":78,"licenseFileUrl":6,"sourceHtmlUrl":92,"categories":93,"keywords":94,"publishedAt":19},"01M7BYB1XSZHKHD8JZTQJ62JN0","Vercel Workflow SDK expert guidance. Use when building durable workflows, long-running tasks, API routes or agents that need pause\u002Fresume, retries, step-based execution, or crash-safe orchestration with Vercel Workflow.","https:\u002F\u002Fgithub.com\u002FCognitionAI\u002Fdevin-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Fvercel\u002Fskills\u002Fworkflow\u002FSKILL.md",[],[],{"skillNo":96,"name":97,"nameZh":6,"description":98,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":78,"licenseFileUrl":6,"sourceHtmlUrl":99,"categories":100,"keywords":101,"publishedAt":19},"01M7BYASGMK935J33E21N5WTW0","verification","Full-story verification — infers what the user is building, then verifies the complete flow end-to-end: browser → API → data → response. Triggers on dev server start and 'why isn't this working' signals.","https:\u002F\u002Fgithub.com\u002FCognitionAI\u002Fdevin-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Fvercel\u002Fskills\u002Fverification\u002FSKILL.md",[],[],{"skillNo":103,"name":104,"nameZh":6,"description":105,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":78,"licenseFileUrl":6,"sourceHtmlUrl":106,"categories":107,"keywords":108,"publishedAt":19},"01M7BY9WJERKSKEH9HK6E71GW2","vercel-storage","Vercel storage expert guidance — Blob, Global Config (formerly Edge Config), and Marketplace storage (Neon Postgres, Upstash Redis). Use when choosing, configuring, or using data storage with Vercel applications.","https:\u002F\u002Fgithub.com\u002FCognitionAI\u002Fdevin-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Fvercel\u002Fskills\u002Fvercel-storage\u002FSKILL.md",[],[],{"skillNo":110,"name":111,"nameZh":6,"description":112,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":78,"licenseFileUrl":6,"sourceHtmlUrl":113,"categories":114,"keywords":115,"publishedAt":19},"01M7BY9N034894YV2J9DCRZHYS","vercel-services","Configure and troubleshoot Vercel Services for multiple frontends and backends in one project. Use when composing a polyglot or multi-service application on one Vercel deployment; defining the `services` key, service-targeted rewrites, or service bindings in `vercel.json`; or running all services with `vercel dev`.","https:\u002F\u002Fgithub.com\u002FCognitionAI\u002Fdevin-marketplace\u002Fblob\u002Fmain\u002Fplugins\u002Fvercel\u002Fskills\u002Fvercel-services\u002FSKILL.md",[],[]]