Skip to content

beifeng724/personal-workbench-skills / skills/personal-workbench-explorer

personal-workbench-explorer

Verified

Explicit-only discovery skill that analyzes the current task context and proposes 3-5 valuable personal or professional workbench ideas, comparing value, usage frequency, build difficulty, data risk, and maintenance cost. Run only when the user explicitly types $personal-workbench-explorer, /personal-workbench-explorer, or names this skill. Never invoke implicitly for general productivity advice, brainstorming, or ordinary workbench questions.

FileKindSize
LICENSEother390 B
SKILL.mdskill md2.2 KB
agents/openai.yamlconfig348 B
references/brief-template.mdmarkdown939 B
references/discovery-workflow.mdmarkdown1.8 KB
scripts/rank_candidates.pyscript1.7 KB

references/discovery-workflow.md

View on GitHub
# Discovery Workflow

## Evidence map

Extract evidence into three columns:

- `stated`: the user directly said it.
- `observed`: present in artifacts or the current task context.
- `hypothesis`: plausible but unconfirmed.

Look for repeated cycles, re-entry of the same data, recurring decisions, deadline tracking, scattered files, status reporting, handoffs, and information that becomes more useful over time.

## Candidate threshold

A candidate workbench should normally satisfy all three:

1. The workflow repeats or accumulates.
2. At least two connected modules or views share information.
3. State needs to persist beyond one session.

A one-off conversion, content-only homepage, or general information question is not a workbench.

## Questions

Ask no more than three at once. Good discriminators include:

- Which activity repeats most often and consumes the most attention?
- What information do you repeatedly search for or re-enter?
- Is this only for you, or must other people see or change it?
- Should changes appear automatically on all your devices?
- Which data would be harmful if exposed?
- How much recurring maintenance or monthly cost is acceptable?

## Scoring

Use a 1-5 scale:

- Value: time saved, mistakes avoided, or decisions improved.
- Frequency: how often the workflow occurs.
- Difficulty: implementation and integration complexity; 5 is hardest.
- Data risk: harm and compliance exposure; 5 is highest.
- Maintenance: ongoing updates, monitoring, and service ownership; 5 is highest.

Do not calculate a misleading single winner from these scores. Explain trade-offs and user fit.

## Output

Use the brief template. Include 3-5 distinct candidates, evidence, assumptions, risks, and the next decision. Finish by telling the user they may invoke `$personal-workbench-builder` with a chosen candidate. Do not invoke it automatically.

reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall