pm-okr

name	pm-okr
description	Manage the OKRs section of `.claude/pm-config.md`. Show current objectives, set new ones, or get AI-suggested updates based on recent work. OKRs drive `/pm` ranking. Use this whenever you want to view, set, or update project OKRs, objectives, key results, or goals. Triggers on "pm okr", "set okrs", "show okrs", "suggest okrs", "project goals", "objectives".
argument-hint	show | set <objectives> | suggest (default: show)

Manage the OKRs (Objectives and Key Results) section of .claude/pm-config.md. OKRs inform /pm — keeping them current ensures prioritization reflects actual goals.

Parse $ARGUMENTS to determine mode:

• No argument or "show": Display current OKRs
• "set ...": Replace OKRs section with the provided text
• "suggest": Analyze recent work and propose OKR updates

Step 1: Verify config exists

test -f .claude/pm-config.md || echo "NO_CONFIG"

If missing, tell the user: "No PM config found. Run /pm first to bootstrap it." Then stop.

Step 2: Parse the config

Extract the ## OKRs section body via the shared parser:

# rc=0 → present with non-empty body; rc=1 → missing or body empty; rc=2 → no config file.
OKRS_CONTENT="$(.claude/scripts/pm-config-get.sh --section OKRs 2>/dev/null)"
OKRS_RC=$?

pm-config-get.sh handles the line-anchored ^## OKRs match (no mid-line matches) and the "next ^## or EOF" boundary. See .claude/scripts/pm-config-get.sh --help.

Mode: show (default)

Display the current OKRs section. If the section is empty or contains only the default placeholder ("No OKRs set"), tell the user:

No OKRs are currently set for this project.

To set OKRs, run: /pm-okr set <your objectives>

Example:
/pm-okr set O1: Launch MVP by April 15
  KR1: All 5 core API endpoints deployed and tested
  KR2: Frontend covers seed selection, pipeline trigger, and results view
  KR3: End-to-end pipeline runs without manual intervention

If OKRs exist, display them formatted clearly with any progress indicators the user has added.

Cross-reference with open issues

After displaying OKRs, fetch the open issue backlog and show which issues align with each key result:

gh issue list --state open --json number,title,labels,body --limit 500

Truncation check: If the returned issue count equals 500, warn: "Showing 500 issues — repo may have more. Results may be incomplete."

For each objective and key result, scan open issue titles, labels, and bodies to find alignment. Match by comparing key terms from the key result text against issue titles, labels, and body content. Only list issues with a clear keyword connection — do not force-match tangential issues.

Display a compact cross-reference below each key result. Sort aligned issues by ascending issue number:

O1: Launch MVP by April 15
  KR1: All 5 core API endpoints deployed — Aligned issues: #42, #45
  KR2: Frontend covers core views — Aligned issues: #50, #52, #55
  KR3: End-to-end pipeline runs — No aligned issues found

If there are no open issues in the repo, skip the cross-reference and note: "No open issues to cross-reference."

Mode: set

Replace the ## OKRs section content with the text provided after "set". Preserve all other sections of pm-config.md verbatim.

1. Parse everything after "set " in $ARGUMENTS as the new OKRs content
2. Read the full .claude/pm-config.md
3. Replace only the content between line-anchored headers: ^## OKRs and the next ^## header
4. Write back the file
5. Confirm: "OKRs updated in .claude/pm-config.md. These will be used by /pm for alignment scoring."

Display the new OKRs for confirmation.

Mode: suggest

Analyze recent repo activity and suggest OKR updates or new OKRs.

Step 3a: Gather recent activity

# Merged PRs in last 30 days
gh pr list --state merged --search "merged:>$(date -v-30d '+%Y-%m-%d' 2>/dev/null || date -d '30 days ago' '+%Y-%m-%d')" --json number,title,body --limit 50

# Closed issues in last 30 days
gh issue list --state closed --search "closed:>$(date -v-30d '+%Y-%m-%d' 2>/dev/null || date -d '30 days ago' '+%Y-%m-%d')" --json number,title --limit 50

# Open issues (to see what's left)
gh issue list --state open --json number,title,labels --limit 500

Truncation check: If the returned issue count equals 500, warn: "Showing 500 issues — repo may have more. Results may be incomplete."

Step 3b: Identify themes

Group merged PRs and closed issues into business themes (same approach as /standup):

• What capabilities were added?
• What areas saw the most investment?
• What's left in the open backlog?

Step 3c: Generate suggestions

If OKRs already exist:

• For each existing objective, assess progress based on completed work — reference both merged PRs and closed issues as evidence (e.g., "KR2 appears 80% complete — 4 of 5 endpoints deployed based on PRs #42, #45, #48 and issue #39 closed as resolved")
• Suggest updating key result progress with specific PR/issue citations
• Flag objectives that may be fully achieved
• Suggest new objectives if the backlog reveals work not covered by current OKRs

If no OKRs exist:

• Propose 2-3 objectives derived from the themes of recent and planned work
• Each objective should have 2-4 measurable key results
• Frame objectives around business outcomes, not technical tasks

Step 3d: Output suggestions

Present suggestions clearly, distinguishing between updates to existing OKRs and new proposed OKRs:

## Suggested OKR Updates

### Existing OKRs — Progress Assessment
- O1: Launch MVP by April 15
  - KR1: All 5 core API endpoints deployed ✅ (completed — PRs #42, #45, #48; issues #30, #33 closed)
  - KR2: Frontend covers core views — 80% (3 of 4 views merged via PRs #50, #51; issue #39 closed; filtering PR #52 still open)
  - KR3: End-to-end pipeline — not started

### Suggested New Objective
- O2: Improve data pipeline reliability
  - KR1: Add retry logic to all scraping stages (3 open issues relate to this)
  - KR2: Reduce pipeline failure rate to <5%

To apply these updates, run: /pm-okr set <paste updated OKRs>

Do not auto-apply suggestions. The user reviews and applies via /pm-okr set.