team_profile ยท diff
git:20260902.af2a47c to git:20260906.f9cef19
35 added, 24 removed. Audit A to A.
---
name: team_profile
- description: Performs deep-dive due diligence on a startup's leadership. Identifies founders, reconciles resumes with LinkedIn, and flags legal/background documents.
+ description: Synthesize a startup team assessment from standard person profiles and collective resume/CV evidence. Use for the original team workflow; team_profile_revised provides the checklist-based assessment.
---
- # Team Profiling Skill
+ # Team profile
- This skill executes a multi-stage reconnaissance and evaluation pipeline for a given startup.
+ Assess leadership and team composition from individual profiles and dataset evidence.
- ## Workflow
+ ## Inputs and outputs
- 1. **Dataset Preparation:**
- * Convert `startup_name` to a dataset slug with `slugify(...)`.
- * Resolve and prepare the startup dataset with `ensure_startup_dataset(...)`.
- * Run `sync_datasets([dataset_slug], raise_on_error=True)` so parsed Markdown and Qdrant are current before analysis.
+ The async `team_profile(startup_name)` returns a one-element `list[InsightFile]`
+ containing the saved or reusable team assessment.
- 2. **Configuration & Insight Cache:**
- * Load `resume_queries`, `team_assessment_prompt`, and optional `linkedin_classification_prompt` from `load_repository_config("team_profile")`.
- * Build an output insight with `lib.insights.InsightFile(dataset=dataset_slug, skill="team_profile", model=llm_model(), config_key=<resume_queries_and_prompts>)`.
- * Use `insight.find(selection="reusable")` and `insight.content()` to reuse a fresh existing team profile when available.
+ ## Workflow and dependencies
- 3. **Person Discovery & Profile Reuse:**
- * Call `person_profile(startup_name, names=None)` to discover and synthesize profiles for the startup's associated people.
- * `person_profile` handles LinkedIn resolution, cached LinkedIn payloads, data-room mentions, personal documents, and generated person-profile insights.
+ Prepare the startup through `ensure_startup_dataset`, synchronize it and check
+ for a reusable team report. On a cache miss, call
+ `person_profile_as_person_objects(startup_name, names=None)` for the existing
+ roster. This uses the [standard person-profile workflow](../person_profile/SKILL.md)
+ and never discovers people. A missing roster must be created separately.
- 4. **Team Context Assembly:**
- * Run `dataset_search(dataset_name=dataset_slug, query=resume_queries)` to collect broader resume/CV/team-document chunks.
- * Deduplicate all person-profile mentions and resume-query chunks by `chunk_id`.
- * Build a single context containing aggregated data-room mentions and discovered person profile summaries.
+ Retrieve collective resume/CV chunks, deduplicate them and individual mentions
+ by `chunk_id`, then synthesize the combined evidence and profile summaries.
+ The registry declares `startup-profile` and `person-profile` prerequisites;
+ the direct workflow calls person profiling only after its team-cache check.
- 5. **LLM Assessment & Output:**
- * Call `generate_markdown(<assembled_context_and_team_profile_instructions>)`.
- * Save the Markdown report with `insight.save(report_md)`, log
- `insight.path`, and return `[insight]`.
+ ## Side effects and failure behavior
+ Startup preparation may import configured Dealum data; synchronization updates
+ derived dataset content. Person profiling may enrich LinkedIn and save profiles.
+ An empty roster permits assessment from the remaining dataset evidence.
+
+ Profile failures propagate. Collective resume-search failures are logged and
+ assessment continues with available context. Generation failures raise.
+ The team cache tracks its own prompts and source revisions, but does not hash
+ supplied person-profile content; editing a profile alone need not refresh it.
+
## Usage
```bash
- conda run -n sictic-env python -m skills.harness /team_profile "<STARTUP_NAME>"
+ conda run -n sictic-env python -m skills.harness /team_profile "<STARTUP>"
```
+
+ The direct CLI uses `--startup`.
+
+ ## References
+
+ - [Implementation](team_profile.py) and [prompts/queries](../../config/team_profile/)
+ - [Bulk registry](../skill_registry.py)
+ - [Checklist workflow](../team_profile_revised/SKILL.md)