Daily Workflows: Daily Project Pulse

Prompt: Daily Project Pulse

Description: A 24-hour change report across your active retained searches, surfacing pipeline movement, stage changes, stalls, and items needing attention.
Frequency: Daily (suggested 7:30am local)
Connectors: Clockwork Recruiting

Prompt: (copy/paste the prompt below into your chat in your LLM)

Using Clockwork Recruiting, generate my Daily Project Pulse for the last 24 hours.

Session setup (do once, cache for this session): (0) Call clockwork_get_knowledge(topic='business-rules') and clockwork_get_knowledge(topic='domain') if not already done this session, and resolve my identity via clockwork_get_current_user. Then fetch and cache: (a) firm users via list_reference_data(data_type=users) — keep only Partner, Recruiter, Researcher, Admin roles; build a userId → name map for note attribution; (b) statuses via list_reference_data(data_type=statuses) → statusId → {name, category, rank} map; (c) note categories via list_reference_data(data_type=note_categories) → categoryId → name map.

Scope: Pull my active retained searches in one call — list_projects(user_id=<me>, status=active, project_type="Retained", include=lead,owner,client_company). Filter out isInternal = true projects.

For each project, output using the <Company> \ <Project> display format:

<Company> \ <Project> — current owner/lead
Candidacies updated in last 24h (name, status.name as current stage, who last touched it) — filter status.rank < 100 to exclude references and benchmarks
Stage changes in the period — for any candidacy updated recently, call get_candidacy_status_history and report concrete moves ("Screened → Client Interview, moved by [user] yesterday"). This is now exact, not inferred.
Candidates removed from consideration (status.category starts with out-) updated in the period (name, reason if captured in notes)
New candidacy notes (1-line summary each, author from cached map, timestamp) via get_candidacy_notes with a 24-hour begin_time/end_time window and include=author
Stall flag: mark any project where no candidacy has been updated in 7+ days
Client visibility: note each updated candidacy's stoplightStatus so you know what's shareable

End with a "Needs My Attention" section listing the top 3 items to act on today and why. When ranking urgency, candidates with higher status.rank (farther along) typically warrant attention first.

Output format: Markdown, grouped by project using <Company> \ <Project> headers, sorted by stall risk (highest first).

Adapt It: Narrow to a single client or search; widen the window to 48–72 hours for a post-weekend catch-up Add a financials line per search (get_project_fee) to keep revenue in view Restrict to searches in client-interview stage and beyond (status.rank >= 1300) for an executive-only version Have it post the summary to Slack instead of returning it inline