thesis-tracker

name: thesis-tracker description: Maintain and update investment theses for portfolio positions and watchlist names. Track key data points, catalysts, and thesis milestones over time. Triggers on "create thesis for [ticker]", "update thesis for [ticker]", "thesis health check [ticker]", "is my thesis still intact", "add catalyst for [ticker]", or "review my positions".

Thesis Tracker

Core question: "Is my original buy reason still valid?"

Workflow

Step 1: Define or Load Thesis

If creating a new thesis:

• Company: Ticker (must exist in REST API — GET /api/companies/{TICKER})
• Position: Long or Short
• Thesis statement: 1-2 sentence core thesis — must be falsifiable
• Key pillars: 3-5 buy reasons with title and description
• Assumptions: 3-5 weighted assumptions with KPI metrics and thresholds (weights must sum to 100%)
• Sell conditions: Specific, actionable exit triggers
• Key risks: 3-5 risks that would invalidate the thesis
• Target price / Stop-loss: Optional price targets

If GET /api/companies/{TICKER} returns 404, tell the user:

I don't have financial data for {TICKER} yet. Please run ETL first:

uv run python -m pfs.etl.pipeline ingest {TICKER} --years 5

This will pull SEC filings, financial statements, and price history. Once done, I can create the thesis.

Do not run ETL yourself — that belongs to the Data Plane.

Optionally seed from company-profile artifacts (--from-profile).

uv run python skills/thesis-tracker/scripts/thesis_cli.py create {TICKER} --interactive

Step 2: Update Log

For each new data point or development (earnings, news, competitor moves, macro shifts):

• Event: What happened and when
• Assumption impacts: Score each assumption (✓ strengthened / ⚠️ weakened / ✗ broken / — no change)
• Thesis strength: Strengthened / Weakened / Unchanged
• Action: Hold / Add / Trim / Exit
• Conviction: High / Medium / Low

uv run python skills/thesis-tracker/scripts/thesis_cli.py update {TICKER} --interactive

Step 3: Health Check

Composite Score = Objective (60%) + Subjective (40%), each 0-100.

• Objective: Weighted KPI scores from REST API financial data
• Subjective: Agent qualitative evaluation (see references/health-check-prompt.md)

Recommendation: ≥75 Hold (strong) · 50-74 Hold (cautious) · 30-49 Trim · <30 Exit

See references/scoring-methodology.md for full methodology.

uv run python skills/thesis-tracker/scripts/thesis_cli.py check {TICKER}

Step 4: Catalyst Calendar

Track upcoming events that could prove or disprove the thesis:

uv run python skills/thesis-tracker/scripts/thesis_cli.py catalyst {TICKER} --add
uv run python skills/thesis-tracker/scripts/thesis_cli.py catalyst {TICKER} --list
uv run python skills/thesis-tracker/scripts/thesis_cli.py catalyst {TICKER} --resolve 1

Step 5: Output

All artifacts go to data/artifacts/{TICKER}/thesis/:

The CLI auto-regenerates the markdown report after every subcommand. Persist via POST /api/analysis/reports.

References

Important Notes

• A thesis must be falsifiable — if nothing could disprove it, it's not a thesis
• Track disconfirming evidence as rigorously as confirming evidence
• Never overwrite update or health check history — always append
• Review theses at least quarterly
• Assumption weights must sum to 100% (auto-normalized if they don't)