Asana Task Analyzer

Asana Task Analyzer

Purpose

Analyze Asana tasks and provide crystal clear implementation guidance. If ANYTHING is unclear or requires assumptions, immediately flag it.

Step 1: Extract Task ID from URL

Asana URLs formats:

• https://app.asana.com/0/{project_id}/{task_id}
• https://app.asana.com/0/{project_id}/{task_id}/f
• https://app.asana.com/0/0/{task_id}/f

Extract the task_id (the last numeric segment before /f).

Step 2: Gather All Task Information

ALWAYS fetch ALL of these using MCP tools:

# 1. Get task details (name, description, assignee, due date, custom fields)
mcp__asana__asana_get_task(
  task_id: "{TASK_ID}",
  opt_fields: "name,notes,html_notes,assignee,assignee.name,due_on,due_at,start_on,completed,projects,projects.name,custom_fields,custom_fields.name,custom_fields.display_value,tags,tags.name,parent,parent.name,dependencies,dependents"
)

# 2. Get all comments and activity (important context!)
mcp__asana__asana_get_stories_for_task(
  task_id: "{TASK_ID}",
  opt_fields: "text,html_text,created_by,created_by.name,created_at,type,resource_subtype"
)

# 3. Get attachments (may contain specs, mockups, designs)
mcp__asana__asana_get_attachments_for_object(
  parent: "{TASK_ID}",
  opt_fields: "name,download_url,view_url,resource_type"
)

Step 3: Analyze and Present

Structure your response as:

## 📋 Task: {task_name}

**Project:** {project_name}
**Due:** {due_date}
**Assignee:** {assignee}
**Status:** {completed ? "Done" : "Open"}

---

## 🎯 Co se po tobě chce

{Clear, actionable description of what needs to be implemented}

---

## 📝 Detailní požadavky

{Bullet points of specific requirements extracted from description and comments}

---

## 💬 Důležité z komentářů

{Key information from comments that affects implementation}

---

## 📎 Přílohy

{List attachments with links - these may contain mockups/specs!}

---

## ⚠️ NEJASNOSTI A DOMNĚNKY

{List ANYTHING that is:
- Not explicitly stated
- Could be interpreted multiple ways
- Requires clarification from task creator
- You had to assume

If nothing is unclear, state: "Zadání je jasné, žádné domněnky."}

---

## ✅ Akční kroky

1. {First concrete step}
2. {Second concrete step}
3. ...

CRITICAL Rules

1. NEVER assume - If something isn't explicitly stated, LIST IT as unclear
2. Read ALL comments - Often contain crucial clarifications
3. Check attachments - May have mockups, designs, or specs
4. Flag ambiguity - Better to ask than to implement wrong
5. Be specific - "Implementovat funkci X" is bad, "Vytvořit endpoint /api/users který vrací..." is good

Example Unclear Items to Flag

• "Jak má vypadat UI?" (pokud není mockup)
• "Jaké validace jsou potřeba?"
• "Co se má stát při chybě?"
• "Jaké jsou edge cases?"
• "Je potřeba zpětná kompatibilita?"
• "Jaký je očekávaný výkon/zátěž?"

Notes

• Task ID is the numeric part at the end of the URL
• Use Czech for the output (as per examples above)
• Always check html_notes for formatted content
• Comments often have more detail than the main description

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: antstudiocz
• Source: antstudiocz/ant-marketplace
• License: MIT
• Homepage: https://antstudio.cz

Install and usage instructions live in the source repository linked above.