Agent-Skills.md

Agent Skills: Task: Fetch Jira Issue

Use when user mentions a Jira issue key (e.g., PLAT-123) or needs context from Jira. Retrieves and formats issue details for PRISM agent workflows.

UncategorizedID: resolve-io/.prism/fetch-jira-issue

Repository

resolve-io/.prism
resolve-io
12

Install this agent skill to your local

pnpm dlx add-skill https://github.com/resolve-io/.prism/tree/HEAD/plugins/prism-devtools/skills/fetch-jira-issue

Skill Files

Browse the full folder contents for fetch-jira-issue.

Download Skill

Loading file tree…

plugins/prism-devtools/skills/fetch-jira-issue/SKILL.md

Skill Metadata

Name
fetch-jira-issue
Description
Use when user mentions a Jira issue key (e.g., PLAT-123) or needs context from Jira. Retrieves and formats issue details for PRISM agent workflows.

Task: Fetch Jira Issue

Purpose

Retrieve and format Jira issue details (Epic, Story, Bug, Task) to provide context for PRISM agent workflows.

Quick Start

  1. Detect or receive Jira issue key (e.g., PLAT-123)
  2. Load Jira config from core-config.yaml
  3. Call Jira REST API via curl
  4. Format response for PRISM workflow context
  5. Return structured issue details

When to Use

  • User mentions a Jira issue key (e.g., PLAT-123)
  • Agent needs context for decomposition, validation, or implementation
  • Explicit *jira {issueKey} command

Prerequisites

  • Jira integration enabled in core-config.yaml
  • Valid Jira credentials configured
  • Issue key provided or detected

Process

Step 1: Detect or Request Issue Key

If issue key mentioned in user message:

  • Extract issue key using pattern from config (default: [A-Z]+-\d+)
  • Inform user: "I found reference to {issueKey}. Let me fetch the details..."

If no issue key detected:

  • Ask user: "Great! Let's take a look at that. Do you have a JIRA ticket number so I can get more context?"
  • Wait for user response
  • If provided, proceed with fetch
  • If not provided or user declines, continue without Jira context

Step 2: Read Jira Configuration

Load config values from ../core-config.yaml:

jira:
  baseUrl: {url}
  email: {email}
  token: {token}
  defaultProject: {project}

Validate configuration:

  • Ensure jira.enabled is true
  • Verify all required fields present
  • If missing, inform user and proceed without Jira context

Step 3: Construct API Request

Build Jira REST API URL:

{baseUrl}/rest/api/3/issue/{issueKey}

Prepare authentication:

  • Use Basic Auth with email:token
  • Format: Authorization header with base64(email:token)

For WebFetch, construct authenticated URL:

https://{email}:{token}@{hostname}/rest/api/3/issue/{issueKey}

Step 4: Fetch Issue Details

Use WebFetch tool with extraction prompt:

Extract and format the following information from this Jira issue:

## Issue Identification
- Issue Key and Type (Epic/Story/Bug/Task/etc)
- Status and Priority
- Project and Components

## Content
- Summary (title)
- Description (full text, preserve formatting)
- Acceptance Criteria (if present in description or as custom field)

## People & Dates
- Assignee
- Reporter
- Created and Updated dates
- Due Date (if set)

## Estimation & Progress
- Story Points (if applicable)
- Original Estimate
- Time Spent
- Sprint (if applicable)

## Relationships
- Epic Link (what epic this belongs to)
- Parent Issue (if subtask)
- Child Issues (count and list)
- Linked Issues (Blocks, Blocked by, Relates to, etc.)

## Additional Context
- Labels
- Fix Versions
- Last 3 comments (with author and date)
- Attachments (count and types)

Format as a clear, structured markdown summary optimized for development context.
If any field is missing or empty, note it as "Not specified" rather than omitting.

Step 5: Handle Response

Success (200 OK):

  • Format and display issue summary to user
  • Store issue data in conversation context for agent reference
  • Include clickable link: [{issueKey}]({baseUrl}/browse/{issueKey})
  • Proceed with agent workflow using fetched context

Issue Not Found (404):

  • Display: "Could not find Jira issue {issueKey}. Please verify the issue key."
  • Ask user if they want to:
    • Try a different issue key
    • Search for issues
    • Proceed without Jira context

Permission Denied (403):

  • Display: "Access denied to {issueKey}. This may require different Jira permissions."
  • Suggest:
    • Verify issue key is correct
    • Check Jira permissions for the account
    • Contact Jira administrator
  • Offer to proceed without Jira context

Authentication Failed (401):

  • Display: "Jira authentication failed. Please check credentials in core-config.yaml."
  • Do not proceed with Jira operations
  • Continue agent workflow without Jira context

Network/API Error:

  • Display: "Unable to connect to Jira at this time. Proceeding without issue context."
  • Log error details for troubleshooting
  • Continue agent workflow without Jira context

Step 6: Present Formatted Summary

Display format:

## 📋 {IssueKey}: {Summary}

**Type:** {Type} | **Status:** {Status} | **Priority:** {Priority}
**Assignee:** {Assignee} | **Reporter:** {Reporter}

### Description
{Description}

### Acceptance Criteria
{Acceptance Criteria or "Not specified"}

### Epic Context
{Epic Link and summary, if applicable}

### Related Issues
- Blocks: {list}
- Blocked by: {list}
- Relates to: {list}
- Child Issues: {count} issues

### Estimation
- Story Points: {points}
- Time Spent: {time}

### Additional Details
- Labels: {labels}
- Components: {components}
- Last Updated: {date}

[View in Jira]({link})

Step 7: Context Integration

For different issue types:

Epic:

  • Fetch child issues count
  • Extract epic goals and scope
  • Identify existing stories to avoid duplication

Story:

  • Extract acceptance criteria carefully
  • Note dependencies and blockers
  • Check for technical notes in comments

Bug:

  • Extract reproduction steps from description
  • Review existing comments for additional context
  • Note severity and customer impact

Task:

  • Extract deliverables and requirements
  • Check for related technical documentation
  • Note any dependencies

Step 8: Offer Follow-up Actions

Based on issue type and agent, offer:

  • "Would you like me to fetch the epic details?" (if story has epic link)
  • "Shall I retrieve all child stories?" (if epic has children)
  • "Would you like to see the linked issues?" (if has dependencies)
  • "Ready to proceed with {agent task}?" (to continue workflow)

Error Recovery

Invalid Issue Key Format:

  • Display: "'{input}' doesn't match Jira issue key format (e.g., PLAT-123)"
  • Ask for correct issue key or offer to proceed without

Multiple Issue Keys Detected:

  • List all detected keys
  • Ask user which one to fetch first
  • Offer to fetch all

Jira Configuration Missing:

  • Display: "Jira integration not configured. Proceeding without issue context."
  • Continue agent workflow
  • Do not block on missing Jira config

Agent-Specific Integration

Story Master (sm):

  • When decomposing epic: fetch epic + all existing child stories
  • When creating story: check for duplicate summaries in epic
  • When estimating: review similar completed stories (if available)

Product Owner (po):

  • When validating story: fetch full acceptance criteria
  • When refining: check for missing fields or incomplete descriptions
  • When prioritizing: review linked dependencies

Support (support):

  • When investigating bug: fetch full reproduction steps and comments
  • When creating tasks: link to original bug report
  • When escalating: include all bug context

QA (qa):

  • When creating tests: fetch acceptance criteria and edge cases
  • When validating: check test requirements from story
  • When reporting: link test results to original story

Dev (dev):

  • When implementing: fetch technical notes from comments
  • When fixing bug: review reproduction steps
  • When completing: verify acceptance criteria met

Architect (architect):

  • When reviewing epic: fetch scope and technical requirements
  • When designing: review architectural decisions in comments
  • When approving: check component relationships

Peer (peer):

  • When reviewing: fetch story context and acceptance criteria
  • When approving: verify implementation matches requirements
  • When suggesting changes: reference architectural constraints

Best Practices

  1. Always ask first if context is ambiguous
  2. Cache fetched data for the conversation session
  3. Format consistently for readability
  4. Handle errors gracefully - never block workflow on Jira failures
  5. Respect privacy - only fetch explicitly referenced issues
  6. Link issues - always include clickable Jira links
  7. Extract acceptance criteria carefully - critical for implementation
  8. Note blocking issues - important for planning
  9. Preserve formatting - maintain description structure
  10. Offer follow-ups - suggest related fetches when relevant

Output

On Success:

  • Formatted issue summary displayed to user
  • Issue context available for agent workflow
  • Clickable Jira link included
  • Follow-up actions offered

On Failure:

  • Clear error message
  • Recovery options offered
  • Workflow continues without Jira context
  • No blocking behavior

Completion Criteria

  • Issue details retrieved or error handled gracefully
  • Formatted summary displayed to user (if successful)
  • Issue context integrated into agent workflow
  • User can proceed with requested task
  • No workflow blocking on Jira failures