Superpowers Skill Bootstrap
Ensures applicable skills are discovered and invoked before every response or action.
The evidence requires checking and invoking applicable skills before any response or action and includes limited platform tool mappings. However, it lacks least-privilege guidance, user confirmation, sensitive-data handling, external-effect disclosure, and rollback requirements, while the “every conversation” trigger is overly broad. No malware, credential theft, covert exfiltration, or other stated red-line behavior is evident, so the score is reduced but not zero.
The rules and platform-adaptation notes are broadly self-consistent, with mappings for Codex, Pi, and Antigravity. However, the key path depends on external skills, runtime tools, and configuration not supplied here, and it lacks handling for unavailable skills, abnormal inputs, and diagnosable failure feedback. Because this is static review without execution evidence, the score remains conservatively below the static ceiling.
The intended audience is coding-agent users and the startup skill-routing scenario is explicit, but the trigger is excessively broad and non-coding boundaries, capability limits, and non-fit cases are unclear. No Chinese-language support or mainland-China network reachability is documented, so points are deducted.
The SKILL.md is readable, its name matches its stated purpose, and referenced files document platform differences. However, the skill itself lacks installation notes, parameters, examples, FAQs, limitations, versioning, changelog, and maintenance ownership; much of that context appears only in the README. Publisher identity is also unverified by the FollowSkills enterprise registry, so points are deducted.
The skill can establish a consistent skill-priority and startup-routing process and may reduce ad hoc agent behavior. Its output is a routing constraint rather than the user’s task result, and its value depends on other skills and host tools. No representative directly usable result or independent execution evidence for this skill’s key path is provided, so the static score is conservative.
The source and platform mappings are auditable, and the repository includes some related integration-test material. However, the supplied tests mainly validate the Antigravity mapping and other components, not the complete using-superpowers trigger, invocation, or failure paths; there is no corroborating independent reproduction evidence, so only limited credit is warranted.
- The “every conversation” rule may cause unnecessary or incorrect invocations; deployments should add precise semantic triggers and skip conditions.
- The required tools, other skills, and host configuration are not fully guaranteed by the supplied material, so failures may shift debugging cost to the user.
- Sensitive-data handling, external side effects, subagent permissions, and recovery are unspecified; the host should add confirmation, isolation, and rollback controls.
- No evidence is provided for Chinese-language support or reachability from mainland-China networks.
What does this skill do, and when should you use it?
Using-superpowers is the bootstrap skill in the obra/superpowers repository. It establishes a mandatory skill-checking rule before responses or actions, including clarification, codebase exploration, and file checks. Before entering plan mode, it requires brainstorming first when brainstorming has not already happened. It does not itself provide coding, testing, or review procedures; it serves as the entry point for the skill system.
Requires the agent to check for applicable skills at conversation start; requires relevant or requested skills to be invoked before any response or action; checks whether brainstorming must occur before plan mode; prioritizes process skills before implementation skills when several apply; requires announcing the skill being used and its purpose; skips itself when the agent was dispatched as a subagent for a specific task.
- Developers using the Superpowers collection who want every conversation to begin with a consistent skill-invocation workflow.
- Teams that want the agent to identify applicable guidance before answering clarifying questions or inspecting a codebase.
- Users preparing an implementation plan who want brainstorming to precede planning when needed.
- Development tasks involving multiple process and implementation skills that require an explicit priority order.
What are this skill's strengths and limitations?
- Covers early actions including responses, clarification, codebase exploration, and file checks.
- Defines a clear priority order for multiple applicable skills.
- Includes a subagent exception to avoid unnecessary bootstrap behavior.
- Distributed as part of an MIT-licensed collection.
- It governs skill discovery and invocation; it does not directly implement coding, testing, or review.
- Its behavior depends on the host environment supporting skill invocation.
- The source provides no independent test results or standalone installation flow for this skill.
- The platform-adaptation section refers to files such as `references/codex-tools.md` that are not included in the supplied source, so their contents cannot be verified here.
How do you install this skill?
The skill has no documented standalone installation command; it is located at skills/using-superpowers/SKILL.md in the Superpowers repository. To install the collection in the Codex app, open Plugins, search for Superpowers, and install it. In Codex CLI, run /plugins, search for superpowers, and select Install Plugin. Other harnesses require their documented collection-specific installation method.
How do you use this skill?
After installing Superpowers, start a conversation; this skill is intended to activate from the beginning. For example, after the prompt “Let's build X,” the agent should first check for and invoke applicable skills. If it is about to enter plan mode without prior brainstorming, it should invoke brainstorming first. The source does not document a standalone manual invocation procedure.