Create long-running engineering plans with scope fidelity and evidence-backed proof gates.
npx skills add https://github.com/rdleclerc/agentic-uber-skills --skill uberplan
Please help me install this Agent Skill. Skill: uberplan Source: https://github.com/rdleclerc/agentic-uber-skills/tree/main/uberplan Command: npx skills add https://github.com/rdleclerc/agentic-uber-skills --skill uberplan
uberplan turns ambiguous, long-running engineering work into a safe, verifiable plan that prevents scope drift, symptom-patching, and false completion.
Use uberplan to plan a long-running coding goal with scope fidelity, a Definition of Done, user expectation/surprise assessment, and evidence-backed proof gates for completion.
High-intent search queries and answers about installing and using this skill.
Scope fidelity prevents scope drift in agentic workflows by requiring a dedicated Scope Fidelity block, blocking overbroad or under-evidenced plans through a confidence gate before execution begins.
An Operational Outcome Contract defines allowed terminal states like operational, blocked, or re_scoped_with_approval. You need it for refactoring and production-risk work to prevent symptom-patching and false completion.
Testing adaptation gates and proof ladders verify completion through an Evidence/Black-box proof pattern, establishing checkpoint cadence and a durable Definition of Done so plans require evidence at each gate.
Yes, this planning approach suits agentic systems and production-risk work by adding adversarial review lanes including an Agent Advocate RCA, Architecture Steward, and First-Principles Simplifier to ensure safety.
You should avoid automated planning when work must not be auto-sliced by similarity, requiring instead scope fidelity and evidence-backed proof gates to safely convert vague work into verifiable contracts.
nmoralescyberConvert vague ideas into structured Spec documents with goals and acceptance criteria.
SollanSystemsScore agent loops against a verifiable contract and gap checklist.
8GG-GitOrchestrate multi-agent PR reviews with adversarial filtering and confidence-rated findings.
PacificStudioConvert clarified specs into OpenASE tickets with dependency edges.
