odin-plan

name	odin-plan
description	Analyzes a TEN JIRA issue and posts an implementation plan as a comment. Does not implement anything — planning only.

Odin Plan — Implementation Planning Skill

Takes a TEN JIRA issue key (e.g. TEN-742), analyzes the problem, determines which repositories and code areas need changes, and posts an implementation plan as a JIRA comment. Does not implement anything or create PRs.

Safety rules — ALWAYS follow these

1. Project scope: Only operate on project TEN. Never create or modify issues in other projects.
2. Planning only: Never create branches, modify code, create PRs, or trigger deployments.
3. Always use --yes on acli edit, transition, and assign commands to avoid interactive prompts.
4. Be concise in JIRA comments. No filler text.

ADF (Atlassian Document Format)

Always use ADF JSON for JIRA comments. Always use --body-file to pass ADF — never pass ADF inline via --body (shell escaping will mangle the JSON and acli will post it as raw text).

Workflow for ADF comments

1. Build the ADF JSON object in memory.
2. Write it to a temp file (e.g. /tmp/odin-comment.json).
3. Pass the file via --body-file.
4. Delete the temp file.

cat > /tmp/odin-comment.json << 'ENDOFJSON'
{"version":1,"type":"doc","content":[{"type":"paragraph","content":[{"type":"text","text":"Comment text"}]}]}
ENDOFJSON
acli jira workitem comment create --key TEN-123 --body-file /tmp/odin-comment.json
rm -f /tmp/odin-comment.json

ADF structure

{"version":1,"type":"doc","content":[...nodes]}

Common node types

Paragraph:

{"type":"paragraph","content":[{"type":"text","text":"Some text"}]}

Heading (level 1-3):

{"type":"heading","attrs":{"level":3},"content":[{"type":"text","text":"Title"}]}

Bullet list:

{"type":"bulletList","content":[
  {"type":"listItem","content":[{"type":"paragraph","content":[{"type":"text","text":"Item"}]}]}
]}

Bold text:

{"type":"text","text":"bold text","marks":[{"type":"strong"}]}

Link:

{"type":"text","text":"click here","marks":[{"type":"link","attrs":{"href":"https://example.com"}}]}

ADF guidelines

• NEVER pass ADF inline via --body — always write to a temp file and use --body-file.
• Keep ADF as single-line JSON (no pretty-printing) inside the file.
• Use heredoc with single-quoted delimiter (<< 'ENDOFJSON') to prevent shell variable expansion.

Command reference (acli)

View an issue

acli jira workitem view TEN-123
acli jira workitem view TEN-123 --json
acli jira workitem view TEN-123 --fields "summary,status,comment,assignee,reporter,description"

Add a comment

cat > /tmp/odin-comment.json << 'ENDOFJSON'
{"version":1,"type":"doc","content":[{"type":"paragraph","content":[{"type":"text","text":"Comment text"}]}]}
ENDOFJSON
acli jira workitem comment create --key TEN-123 --body-file /tmp/odin-comment.json
rm -f /tmp/odin-comment.json

List and delete comments

# List comments as JSON (for finding comment IDs)
acli jira workitem comment list --key TEN-123 --json --paginate

# Delete a comment by ID
acli jira workitem comment delete --key TEN-123 --id <comment-id>

Steps

When invoked with a JIRA issue key (e.g. /odin-plan TEN-742):

Step 1 — Assign and transition the issue

Assign the issue to the person who invoked the skill and move it to "Under arbeid":

acli jira workitem assign --key TEN-742 --assignee "@me" --yes
acli jira workitem transition --key TEN-742 --status "Under arbeid" --yes

Step 2 — Analyze the JIRA issue

Fetch the full issue details to understand what needs to be done:

acli jira workitem view TEN-742 --json

Read the summary, description, and comments carefully. Understand:

• What is the problem or request?
• What behavior is expected vs actual?
• Are there stack traces, error messages, or specific SED/BUC types mentioned?
• Are there any comments with additional context?

Step 3 — Determine which repositories need changes

Use the EUX Architecture document (navikt/eux-architecture README.md) to understand the platform and identify which repositories are involved.

To read the architecture overview:

cat eux-architecture/README.md

Or if not locally available:

gh api repos/navikt/eux-architecture/contents/README.md --jq '.content' | base64 -d

Based on the issue analysis:

• Map the problem to specific services (e.g. journaling issues → eux-journalfoering, eux-journal; RINA case issues → eux-nav-rinasak, eux-rina-api; frontend issues → eux-web-app, eux-neessi).
• Clone or navigate to the relevant repositories.
• Search the code to find the exact files and functions that need changes.

Step 4 — Build the implementation plan

Create a detailed plan that covers:

1. Problem summary — what the issue is about, in one or two sentences.
2. Affected repositories — which repos need changes and why.
3. Changes per repository — for each repo, list:
   • Which files need to be modified or created
   • What the change should be (describe the logic, not just "fix it")
   • Any new classes, endpoints, or components needed
4. Risks and considerations — edge cases, breaking changes, migration concerns, or things to verify.
5. Dependencies — if changes must be deployed in a specific order.

Be specific enough that someone (or odin-jira) can implement the plan without re-analyzing the issue.

Step 5 — Delete any existing plan comment, then post the new one

There must be at most one "Odin's plan" comment on an issue. Before posting, find and delete any previous plan comment.

5a — Find and delete the existing plan comment

List all comments as JSON and look for one whose body contains the text "Odin's plan":

acli jira workitem comment list --key TEN-742 --json --paginate

Parse the JSON output to find any comment where the rendered body or ADF content contains "Odin's plan". Extract the comment id for each match.

If a matching comment is found, delete it:

acli jira workitem comment delete --key TEN-742 --id <comment-id>

If multiple matching comments exist (shouldn't happen, but be safe), delete all of them so only the new one remains.

5b — Post the new plan comment

Add a comment with the heading "Odin's plan". The plan MUST start with an ADF heading node with text "Odin's plan" so that odin-jira can find it.

Example structure:

cat > /tmp/odin-plan-comment.json << 'ENDOFJSON'
{"version":1,"type":"doc","content":[{"type":"heading","attrs":{"level":2},"content":[{"type":"text","text":"Odin's plan"}]},{"type":"heading","attrs":{"level":3},"content":[{"type":"text","text":"Problem"}]},{"type":"paragraph","content":[{"type":"text","text":"<concise problem summary>"}]},{"type":"heading","attrs":{"level":3},"content":[{"type":"text","text":"Affected repositories"}]},{"type":"bulletList","content":[{"type":"listItem","content":[{"type":"paragraph","content":[{"type":"text","text":"eux-neessi","marks":[{"type":}]},{:,:}]}]}]},{:,:{:3},:[{:,:}]},{:,:{:4},:[{:,:}]},{:,:[{:,:[{:,:[{:,:}]}]}]},{:,:{:3},:[{:,:}]},{:,:[{:,:[{:,:[{:,:}]}]}]}]}
ENDOFJSON
acli jira workitem comment create --key TEN-742 --body-file /tmp/odin-plan-comment.json
 -f /tmp/odin-plan-comment.json

Step 6 — Final check

• Verify the JIRA comment was created successfully.
• If any step failed, comment on the JIRA issue explaining what went wrong.