3p Update

Not all sources are required. The skill produces output from whatever sources are available and notes gaps.

Output Modes

Optional Flags

Pipeline

Step 1: Parse input and determine mode

• If a customer name is provided: per-customer mode (single customer 3P)
• If no customer name: cross-customer mode (portfolio 3P covering all customers in the registry)
• Check for --confluence flag

Step 2: Load customer registry

Read templates/customers.yaml to resolve customer data.

• Per-customer mode: Find the matching customer entry by name (case-insensitive)
• Cross-customer mode: Collect all customers where action_tracker_id is not "PLACEHOLDER"

If per-customer mode and the customer is not found in the registry, tell the user and stop.

Step 3: Gather data from three sources

For each customer in scope, gather data from all available sources. Skip sources that are not configured (PLACEHOLDER GIDs/IDs) and note the gap in the output.

3a. Asana tasks (Progress + Plans)

uv run --project .claude/skills/asana python .claude/skills/asana/scripts/query.py tasks \
  --project-gid <action_tracker_id> --limit 100 --pretty

Process the response:

• Recently completed (last 7 days, completed_at within range): feed into Progress
• In Progress section (active work): feed into Progress
• To Do / Scheduled/Future sections (upcoming work): feed into Plans
• Overdue tasks (past due_on, not completed): feed into Problems
• Stale tasks (In Progress or To Do, modified_at > 7 days ago): feed into Problems

If action_tracker_id is "PLACEHOLDER": skip Asana, note "Asana not configured" in output.

3b. Jira activity (Progress + Problems)

uv run --project .claude/skills/jira python .claude/skills/jira/scripts/issues.py list \
  --customer "<jira_customer>" --max-results 50 --with-comments

Process the response:

• Recently resolved issues (status = Resolved/Closed/Done, updated in last 7 days): feed into Progress
• Issues with recent engineering comments (non-FE-UPDATE comments in last 7 days): feed into Progress
• Stale/blocked issues (no updates 14+ days, still open): feed into Problems
• New P0/P1 issues (created in last 7 days): feed into Problems

Note: FE-UPDATE comments from SEs do NOT count as engineering activity (prevents gaming staleness).

3c. Slack signals (Problems + Progress)

For each channel in the customer's slack_channels where id is not "PLACEHOLDER":

uv run --project .claude/skills/slack python .claude/skills/slack/scripts/channels.py history \
  --channel <channel_id> --limit 50 --oldest <7_days_ago_unix_timestamp>

Process the response:

• Frustration signals (escalation language, mentions of alternatives, repeated asks): feed into Problems
• Positive signals (thanks, appreciation, successful outcomes, shipped features): feed into Progress

If no Slack channels are configured (all IDs are PLACEHOLDER): skip Slack, note "Slack channels not configured" in output.

Step 4: Synthesize 3P update

Read all gathered data and produce the 3P output.

Tone guidance:

• Concise, data-driven, factual
• 1-3 sentences per section
• Reference specific Jira issues (WB-XXXX) and Asana tasks by name
• No filler or generic statements
• If a section has nothing noteworthy, say "No significant updates this period" rather than fabricating content

Per-customer format

[emoji] [CustomerName] -- [date]

**Progress**
[1-3 sentences: what happened this week. Cite Jira issues resolved, Asana tasks completed, positive Slack signals.]

**Plans**
[1-3 sentences: what's coming next. Cite Asana tasks in To Do/In Progress with upcoming due dates.]

**Problems**
[1-3 sentences: what's blocked or at risk. Cite overdue Asana tasks, stale Jira issues, negative Slack signals. If nothing: "No blockers identified."]

Emoji selection:

• Rocket for momentum (lots of progress, things moving fast)
• Construction for heavy work (large active workload)
• Warning for problems (blockers, escalations)
• Checkmark for smooth sailing (no issues, steady progress)

Cross-customer (portfolio) format

[clipboard emoji] SE Portfolio Update -- [date]

[For each customer with data:]

**[CustomerName]**
- Progress: [1 sentence]
- Plans: [1 sentence]
- Problems: [1 sentence or "None"]

[End with:]

**Cross-cutting**
- [Any themes across customers: common blockers, shared wins, resource constraints]

Step 5: Output

Default: Print the 3P text to the user as Slack-ready output (copy-paste into Slack).

If --confluence flag is set:

Convert the 3P text to simple HTML (paragraphs, bold, links) and create a Confluence page:

uv run --project .claude/skills/confluence python .claude/skills/confluence/scripts/pages.py create \
  --space "~PERSONAL_SPACE" --title "3P Update -- [CustomerName] -- [date]" \
  --body "<html-content>"

Print both the text output and the Confluence page URL.

Step 6: Present to user

Show the 3P update text. Include a brief footer noting:

• Data sources used: which sources returned data (Asana, Jira, Slack)
• Data gaps: which sources were skipped and why (not configured, no data, PLACEHOLDER)
• Confluence URL: if --confluence was used, include the page link

Anti-patterns

• Do NOT fabricate data or connections. If a source returns nothing, say so.
• Do NOT include generic filler ("The team continues to make progress..."). Every sentence must reference specific data.
• Do NOT include internal Slack sentiment details in the output (that belongs in the intelligence dashboard via /customer-snapshot).
• Do NOT attempt to read Gong, email, or other sources not listed above.
• Do NOT attribute one customer's data to another customer (in portfolio mode, keep data strictly separated).

Troubleshooting

Data Source Reference

Related Skills

• /asana -- Base skill for Asana task queries and mutations
• /jira -- Base skill for Jira issue queries and FE-UPDATE management
• /slack -- Base skill for Slack channel history and search
• /confluence -- Publishing target for --confluence flag
• /customer-snapshot -- Full intelligence dashboard (deeper analysis than 3P)
• /cadence-prep -- Meeting prep (complementary to 3P -- agenda vs status update)