菜单
Harper Security Review skill owned by Harper (Security Engineering Director). Runs on: security-sensitive PR. Scaffold generated from the Mindshare skills registry; install into ~/.claude/skills/harper-security-review/ to use.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Build role agents for a root organization or multi-agent corporation. Use when the user wants to create an organizational role such as CEO, CFO, advisor, operator, manager, support lead, data steward, research lead, or functional department voice, and needs to decide role mode, engagement type, authority, implementation form, learning behavior, and whether the role should be advisory, workflow-owned, skill-backed, loop-backed, autonomous, human-in-the-loop, or only a lightweight prompt/persona.
Observe CLI model calls, prompt/input payloads, stdout/stderr, provider-reported usage metadata, and local token estimates across Codex, Claude, DeepSeek, OpenAI, Anthropic, or other AI CLIs. Use when a user needs truthful token visibility, prompt-origin accounting, usage audits, or a harness around model CLI input/output; labels exact provider usage separately from local estimates and blocks claims that cannot be proven.
Observe CLI model calls, prompt/input payloads, stdout/stderr, provider-reported usage metadata, and local token estimates across Codex, Claude, DeepSeek, OpenAI, Anthropic, or other AI CLIs. Use when a user needs truthful token visibility, prompt-origin accounting, usage audits, or a harness around model CLI input/output; labels exact provider usage separately from local estimates and blocks claims that cannot be proven.
Create an agent brief for the MAPS Define phase from an existing role. Use when turning an approved role artifact into an agent brief with goals, users, authorization, scope, non-goals, success criteria, failure criteria, escalation points, risks, and assumptions.
Accounting skill owned by Finn (Finance Director). Runs on: scheduled + emails link. Scaffold generated from the Mindshare skills registry; install into ~/.claude/skills/accounting/ to use.
Adrian Legal Gate skill owned by Adrian (Legal Director). Runs on: external commitment / contract. Scaffold generated from the Mindshare skills registry; install into ~/.claude/skills/adrian-legal-gate/ to use.
| name | harper-security-review |
| description | Harper Security Review skill owned by Harper (Security Engineering Director). Runs on: security-sensitive PR. Scaffold generated from the Mindshare skills registry; install into ~/.claude/skills/harper-security-review/ to use. |
Owned by Harper (Security Engineering Director). Generated scaffold from the Mindshare skills registry (roles/skills-registry.json). This file is a real, installable Claude skill: drop the harper-security-review/ folder into ~/.claude/skills/ and invoke it with /harper-security-review. The structure below is honest scaffold depth — the owning role deepens the body with its real procedure.
Provide the harper security review capability for Harper (Security Engineering Director). Use this skill when the owning role needs a repeatable, registered procedure rather than an ad-hoc prompt.
/harper-security-review or when the user describes the task this skill names.Tool ability is not authority. This scaffold defaults to report-only. Any step that commits, sends, spends, or changes control-plane state stays blocked until the gate owner approves.
A reviewable report or artifact for Harper (Security Engineering Director). No silent state changes. Every gated action is surfaced for approval, not auto-applied.
This file grants no authority by itself. Activation, autonomy, git/release, secrets, external comms, and spending all remain gated to their registered owners until explicitly approved.