Finding-Unknowns Skills
Neeeophytee/finding-unknowns-skills
14 skills that help coding agents surface unknowns, clarify requirements and verify bug fixes
Finding-Unknowns Skills is a pack of 14 installable agent skills for clarifying requirements, managing context and testing assumptions. The repository lists Claude Code, OpenAI Codex and Hermes as supported agents.
What Finding-Unknowns Skills does
Finding-Unknowns Skills treats your prompt as a map and the codebase as the territory, and each skill reduces the gap before implementation makes it expensive. Skills for the Before phase include blindspot-pass, brainstorm-prototypes, interview-me, reference-hunt and implementation-plan. Implementation-notes logs deviations from the plan during work, and pitch-packager and change-quiz cover the After phase with a buy-in document and a comprehension quiz to pass before merging.
Three more skills work on the context every prompt inherits, including context-audit, which finds contradictions, duplicates and dead rules in CLAUDE.md files and skills, and agent-interface-design. The GitHub description adds skills for testing assumptions and verifying bug fixes with regression evidence. The skills follow the agentskills.io SKILL.md format, and the repository includes examples and a compatibility document. The README says twelve skills are distilled with attribution from public essays by Thariq Shihipar, plus two maintainer-designed extensions.
Key features
- 14 skills across before, during and after phases of a task
- Question-driven skills such as interview-me and blindspot-pass
- context-audit for cleaning contradictions out of CLAUDE.md files and skills
- Plan, notes and pitch skills that document decisions and deviations
- Portable agentskills.io SKILL.md packaging with examples and compatibility notes
When to use it
- Surfacing missing requirements before asking an agent to build a feature
- Auditing the instruction files your agent loads on every task
- Passing a short comprehension quiz on a change before merging it
Who it is for: Developers who work with coding agents and want prompts and context checked before expensive implementation work.
How it fits with Hermes Agent
The repository description and topics list Hermes among the supported agents, and the skills use the same SKILL.md format Hermes loads.
Note: It is a community project distilled from public essays by Thariq Shihipar and is not an official Anthropic repository.
FAQ
What is Finding-Unknowns Skills?
Finding-Unknowns Skills is a pack of 14 installable agent skills that help a coding agent find unknowns, clarify requirements, manage context, test assumptions and verify bug fixes.
Does Finding-Unknowns Skills work with Hermes Agent?
Yes, the repository lists Hermes among supported agents next to Claude Code and OpenAI Codex. Any agent that reads agentskills.io SKILL.md files should be able to load the skills.
Is Finding-Unknowns Skills free and open source?
Yes. The repository is released under the MIT license. It is a community project and not an official Anthropic repository.
Similar skills for Hermes Agent
All skillsAgent skill that applies Charlie Munger's mental models and decision heuristics to your questions
Minara-AI Minara SkillsTrading skills for agents: stocks, futures, forex and crypto in plain language, with a Hermes setup script
sergebulaev Instagram SkillsNine Instagram marketing skills for Claude Code and Codex, installable into Hermes Agent as SKILL.md files
AaravKashyap12 Advise Project ApproachPortable project-planning skill that checks evidence before recommending a stack or architecture
alchaincyf Karpathy SkillAgent skill that applies Andrej Karpathy's mental models and decision heuristics, with a Chinese README
superzhang21 GhostwriterPersona library that makes an AI write in the voice of named characters, installable as a Hermes skill
Related guides: What is Hermes Agent? · SOUL.md for Hermes Agent: what it is and how to write one