CtrlK
BlogDocsLog inGet started
Tessl Logo

oh-my-ai/discovery-interview

Interactive discovery interview → structured product spec. Triggers: spec, PRD, requirements, scoping, brainstorming, new project. Uses AskUserQuestion; WebSearch/WebFetch when user wants research. Outputs user stories, acceptance criteria, constraints.

99

2.28x
Quality

100%

Does it follow best practices?

Impact

96%

2.28x

Average score across 3 eval scenarios

SecuritybySnyk

Advisory

Suggest reviewing before use

Overview
Quality
Evals
Security
Files

criteria.jsonevals/scenario-1/

{
  "context": "Tests whether the agent conducts a thorough discovery interview with sufficient question depth, includes uncertainty options in every question, performs at least one research loop, and covers all required interview phases before writing the spec.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Minimum question count",
      "description": "The interview_log.md contains at least 10 distinct question exchanges (Q1 through Q10 or more labeled questions)",
      "max_score": 15
    },
    {
      "name": "Uncertainty option present",
      "description": "Every question shown in interview_log.md includes at least one option expressing uncertainty (e.g. 'I'm not sure', 'Research this', 'Not certain', or similar phrasing) among the answer choices",
      "max_score": 15
    },
    {
      "name": "Research loop included",
      "description": "The interview_log.md shows at least one research loop: a question offering research, the user accepting, followed by a search or fetch action, followed by a follow-up question building on that research",
      "max_score": 15
    },
    {
      "name": "All 7 phases present",
      "description": "The interview_log.md shows coverage of all 7 phases: initial orientation, category deep dive, research loops, conflict resolution (or acknowledgment none exist), completeness check, spec generation confirmation, and implementation handoff",
      "max_score": 10
    },
    {
      "name": "Phase 7 handoff options",
      "description": "The interview_log.md includes a final handoff question offering at least two of: 'Start implementation', 'Plan implementation', 'Review spec / done' as distinct options",
      "max_score": 10
    },
    {
      "name": "Spec file date-prefixed name",
      "description": "The spec file is saved under docs/specs/ with a filename starting with a date in YYYY-MM-DD format (e.g., docs/specs/2026-04-06-fitness-app.md)",
      "max_score": 10
    },
    {
      "name": "Spec has required sections",
      "description": "The spec file contains all of: Executive Summary, Problem Statement, Success Criteria, User Personas, User Journey, Functional Requirements (with P0/P1/P2 sub-sections), Technical Architecture, Non-Functional Requirements, Out of Scope",
      "max_score": 15
    },
    {
      "name": "Completeness check before spec",
      "description": "The interview_log.md shows a completeness check step (verifying problem, UX, technical, and decisions areas) occurring before the spec file is written",
      "max_score": 10
    }
  ]
}

evals

scenario-1

criteria.json

task.md

CATEGORIES.md

SKILL.md

SPEC_TEMPLATE.md

tile.json