Task Analysis
Folder: .github/skills/tsh-task-analysing/
Used by: Context Engineer, E2E Engineer
Provides a structured process for gathering and expanding task context from multiple sources.
Process
Step 0: Determine Input Source
Before anything else, determine where the context comes from:
| Source | How to Identify |
|---|---|
| Research/plan files | Check specifications/ folder for .research.md or .plan.md |
| Task ID | Jira ID, GitHub issue number, or similar identifier |
| Prompt context | Free-form description in the chat message |
Step 1: Search External Sources
Discover and search relevant external tools:
- Jira — Task details, subtasks, parent tasks, linked issues.
- Confluence — Knowledge base articles, process documentation.
- Notion, Linear, Trello — Alternative project management tools.
- GitHub — Issues, PRs, discussions.
Step 2: Gather Information
Collect information from all available sources:
- Task management tools (requirements, acceptance criteria).
- Knowledge bases (domain context, business rules).
- Codebase (existing implementation, patterns, constraints).
- Figma designs (UI requirements, user flows).
Step 3: Identify Gaps
Ask clarifying questions for:
- Ambiguous requirements.
- Missing acceptance criteria.
- Undefined edge cases.
- Conflicting information between sources.
Step 4: Produce Research Report
Generate a structured document (research.example.md template) containing all gathered context, ready for the planning phase.
Connected Skills
tsh-codebase-analysing— For understanding the existing codebase in context.tsh-implementation-gap-analysing— For identifying what exists vs what needs to be built.