prathamchopra001/skills/gap_analysis

SKILL.md: gap_analysis

Task Description

Identifies gaps between research objectives and completed work by comparing stated goals against findings and tasks, categorizing answered questions, open questions, and weak areas.

Output Format

JSON Schema:

{
  "answered_questions": ["specific question 1", "specific question 2"],
  "open_questions": ["specific question 3", "specific question 4"],
  "weak_areas": ["area 1", "area 2"]
}

All values are arrays of strings. Use exact phrasing from input.

Decision Framework

1. Parse Objective: Break down the objective into explicit research questions (e.g., "What are the key metrics for biodiversity monitoring?").
2. Cross-Reference Findings: Match completed tasks/findings to answered questions (e.g., "GBIF data integration" → answers "How can GBIF data be leveraged?").
3. Identify Open Questions: Flag questions not addressed by findings (e.g., "What are the ethical implications of real-time surveillance?").
4. Highlight Weak Areas: Note gaps in methodology, data, or scope (e.g., "Lack of field testing for OpenTelemetry deployment").
5. Confidence Scoring: Rate each question's confidence (0–100) based on evidence. Only include questions with ≥70 confidence in "answered_questions".
6. Edge Cases:
   • Ambiguous objectives → classify as weak areas.
   • Incomplete data → flag as weak areas, not open questions.

Common Patterns

✅ Example 1

Input:
Objective: "Investigate how observability engineering principles... can be applied to frontline ecological incident monitoring..."
Output:

{
  "answered_questions": ["What are the key metrics for biodiversity monitoring?"],
  "open_questions": ["How do OpenTelemetry tools handle real-time ecological data streaming?"],
  "weak_areas": ["Lack of case studies on biodiversity surveillance"]
}

✅ Example 2

Input:
Objective: "Evaluate renewable energy adoption in rural communities."
Output:

{
  "answered_questions": ["What are the cost barriers for solar adoption in rural areas?"],
  "open_questions": ["How do policy incentives vary across regions?"],
  "weak_areas": ["Insufficient data on long-term maintenance costs"]
}

Anti-Patterns

• ❌ Rephrasing completed tasks as open questions (e.g., "GBIF data integration" → "What is GBIF data integration?").
• ❌ Using vague entries like "Need more research" or "Unknown".
• ❌ Speculative questions not supported by objective.
• ❌ Duplicating entries across categories.
• ❌ Including tasks as open questions if they’re just uncompleted.

Quality Checklist

1. Validate all answered questions are explicitly addressed in findings/tasks.
2. Ensure open questions are genuinely unanswered, not just unexplored.
3. Flag weak areas for insufficient data/methodology, not missing tasks.
4. Confirm JSON structure matches schema (no extra fields).
5. Remove duplicates and ensure phrasing matches input.