agentleFS
AtmoskunAtmoskun/claude-skills/skills/thesis/SKILL.md

thesis

Skill0 starsChanged 9 months agoMITView on GitHub

Investment Thesis Manager - Create, view, and update investment thesis documents for portfolio positions

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

Show its rules and text

We split this skill into the rules it gives an agent

31 rules. About 0 are common, 17 are distinctive, and 14 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈17

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 14

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Check if "update" subcommand with note

    line 96
  2. Very specific:

    Read and display the thesis file (thesis.md if exists, otherwise summarize thesis.yaml)

    line 116
  3. Distinctive:

    If no kill_criteria in YAML: show "No kill criteria defined.

    line 129
  4. Very specific:

    If user wants details: run python shared/13f_query.py {TICKER} for full markdown table

    line 138
  5. Distinctive:

    If no 13F data available: silently skip (don't show empty section)

    line 139
  6. Distinctive:

    If no sizing fields: show "No sizing data.

    line 150
  7. Distinctive:

    Add thesis log entry

    line 153
  8. Distinctive:

    Update bull/bear case

    line 155
  9. Distinctive:

    Run NLM first-mention query against 投资观点周报 notebook:

    line 163
  10. Distinctive:

    If NLM returns a result, present to user:

    line 171
  11. Distinctive:

    Fallback: If NLM fails or returns no result → proceed directly to manual Info Source selection

    line 177
  12. Distinctive:

    Ask: "What are the core moats, growth drivers, and key threats?"

    line 189
  13. Distinctive:

    Create thesis.md from template (include NLM attribution fields in frontmatter)

    line 194
  14. Distinctive:

    Add entry to Thesis Log table with today's date

    line 201
  15. Very specific:

    Read research/companies/{TICKER}/thesis.yaml — extract kill_criteria list

    line 337
  16. Very specific:

    Quantitative: Search for latest data (web search, earnings analysis in Obsidian, NotebookLM history).

    line 339
  17. Very specific:

    Update thesis.yaml with new check_result, check_note, and last_checked date

    line 341
  18. Very specific:

    Read existing thesis from research/companies/{TICKER}/thesis.md — extract core thesis, bull/bear cases

    line 391
  19. Distinctive:

    Search NLM answer history (free, no daily limit cost):

    line 392
  20. Distinctive:

    Search 13F institutional holdings for smart money signals:

    line 401
  21. Very specific:

    Create directory if needed: ~/PORTFOLIO\research\companies\{TICKER}\

    line 484
  22. Very specific:

    Check if passed.md already exists

    line 485
  23. Distinctive:

    If NEW: Ask 3 quick questions using AskUserQuestion tool:

    line 486
  24. Very specific:

    Create passed.md from template below

    line 498
  25. Very specific:

    If UPDATE (/thesis TICKER passed update): append a new entry to the Passed Log table.

    line 501
  26. Very specific:

    Add a ## Framework Coverage section at the end of the thesis template, after Thesis Log:

    line 650
  27. Distinctive:

    After creating a new thesis (step 4B), automatically run:

    line 679
  28. Distinctive:

    Add framework coverage check after kill criteria check (step 4 in check workflow):

    line 689
  29. Very specific:

    If any section has degraded since last scan → flag it.

    line 693
  30. Very specific:

    If framework_coverage.last_scanned is >30 days ago → suggest re-scan.

    line 694
  31. Very specific:

    When a trade is logged via /trade, the thesis position history should be updated automatically if the thesis exists.

    line 698

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Investment Thesis Manager

Manage investment thesis documents for portfolio positions. Structure matches the Trading Journal web app for conceptual consistency.

Project Location

~/PORTFOLIO

Syntax

/thesis {TICKER}
/thesis {TICKER} update "{note}"
/thesis {TICKER} check
/thesis {TICKER} passed
/thesis {TICKER} passed update

Examples:

  • /thesis AAPL - View or create thesis for AAPL
  • /thesis TSM update "Q4 beat, raising target" - Add thesis log entry
  • /thesis PM check - Check all kill criteria against latest data
  • /thesis NVDA passed - Record why you passed on NVDA (creates passed.md)
  • /thesis NVDA passed update - Update an existing passed record

Classification Fields (Match Web App)

Industry (13 options)

CodeIndustry
BMATBasic Materials
CCYCConsumer Cyclicals
CNCYConsumer Non-Cyclicals
EDUCEducation
ENRGEnergy
FINAFinancials
HLTHHealthcare
INDUIndustrial
REALReal Estate
TECHTechnology
TELETelecomm
UNCLUnclassified
UTILUtilities

Strategy (5 options)

StrategyDescription
ValueUndervalued by market
GrowthHigh growth potential
GARPGrowth at reasonable price
Event-DrivenCatalyst-based
MomentumTrend following

Driver (7 options)

DriverDescription
ValuationPrice vs intrinsic value
GrowthRevenue/earnings growth
MomentumPrice momentum
EventSpecific catalyst
MacroEconomic factors
TechnicalChart patterns
CatalystUpcoming event

Info Source (9 options)

SourceDescription
Self-ResearchOwn analysis
Sell-SideAnalyst reports
Social MediaTwitter/X, etc
NewsNews articles
PodcastInvestment podcasts
13FInstitutional filings
FriendPersonal network
EarningsEarnings call/report
OtherOther source

Planned Hold Period

PeriodDays
30 daysShort-term trade
60 daysNear-term
90 daysQuarter
180 daysHalf year
365 daysLong-term

Workflow

1. Parse Arguments

Extract ticker from command (e.g., /thesis AAPL -> AAPL) Check if "update" subcommand with note

2. Fetch Current Position

curl -s http://localhost:8000/api/portfolio

Extract for this ticker:

  • Current shares
  • Average cost
  • Market value
  • Unrealized P&L
  • % of portfolio

3. Check if Thesis Exists

File location: ~/PORTFOLIO\research\companies\{TICKER}\thesis.md

4A. If Thesis EXISTS: Display It

  1. Read and display the thesis file (thesis.md if exists, otherwise summarize thesis.yaml)
  2. Show current position status from portfolio API
  3. Show Kill Criteria status table (from thesis.yaml kill_criteria field):
    Kill Criteria (5 active)
    | # | Condition                                              | Type | Result  | Checked    |
    |---|--------------------------------------------------------|------|---------|------------|
    | 1 | ZYN US share <50% OR category <10% OR competitor +5pp  | QT   | Pass    | 2026-02-07 |
    | 2 | FDA lowers PMTA barriers                               | QL   | Pass    | 2026-02-07 |
    | 3 | Global SFP category growth <5% for 2Q                  | QT   | Pass    | 2026-02-07 |
    | 4 | IQOS Japan HTU share <65%                              | QT   | Pass    | 2026-02-07 |
    | 5 | 3+ markets impose punitive SFP taxes/bans              | QL   | Pass    | 2026-02-07 |
    • If no kill_criteria in YAML: show "No kill criteria defined. Use /thesis TICKER to add."
    • Flag any criteria with check_result = "warning" or "fail"
    • Flag any criteria where last_checked is >14 days ago (qualitative) or >30 days ago (quantitative)
  4. Show peers if defined in thesis.yaml
  5. Show 13F Institutional Activity (auto, from shared/13f_query.py):
    cd ~/.claude/skills && python shared/13f_query.py {TICKER} --summary
    • Display one-line summary: "{TICKER} (Q3): 12 holders (2 new, 3 up, 1 down). Top: Scion Asset (4.2%)"
    • If user wants details: run python shared/13f_query.py {TICKER} for full markdown table
    • If no 13F data available: silently skip (don't show empty section)
    • For Python import: use importlib.import_module('shared.13f_query')
  6. Show Position Sizing (from thesis.yaml sizing fields):
    📐 Sizing: conviction(4/High) × quality(B/1.0) = 5% × 1.5 × 1.0 = 7.5%
    当前实际仓位: 8.2% → ⚠️ 略超建议 (0.7% over)
    • Formula: target = base_size_pct × conviction_multiplier × quality_multiplier
    • Conviction: 1=0.5x, 2=0.75x, 3=1.0x, 4=1.5x, 5=2.0x
    • Quality: A=1.2x, B=1.0x, C=0.7x
    • Hard cap: min(result, 10%)
    • If no sizing fields: show "No sizing data. conviction={N}, quality={grade}"
    • Get actual position weight from curl -s http://localhost:8000/api/portfolio and compare
  7. Offer options:
    • Add thesis log entry
    • Update catalysts
    • Update bull/bear case
    • Update position history
    • Check kill criteria (/thesis TICKER check)

4B. If Thesis DOES NOT EXIST: Create New (NLM-Assisted)

  1. Create directory: ~/PORTFOLIO\research\companies\{TICKER}\
  2. NLM Attribution Query (auto, before asking user):
    • Run NLM first-mention query against 投资观点周报 notebook:
      cd ~/.claude/skills && python -c "
      from shared.nlm_attribution import query_first_mention
      result = query_first_mention('{TICKER}', '{COMPANY_NAME}')
      print(result)
      "
    • If NLM returns a result, present to user:

      "根据周会纪要,{TICKER} 最早在 {date} 讨论过:'{citation}' 初始情绪: {bullish/bearish/neutral} 确认这是想法来源?还是来自其他渠道?"

    • User confirms → auto-fill idea_source: weekly-meeting, source_detail, nlm_citation, first_seen
    • User overrides → use their selection from Info Source options
    • Fallback: If NLM fails or returns no result → proceed directly to manual Info Source selection
  3. Ask user for:
    • Industry (show options)
    • Strategy (show options)
    • Driver (show options)
    • Info Source (pre-filled if NLM succeeded, show options if not)
    • Core thesis (1-2 sentences)
    • Bull case with target price
    • Bear case with stop price
    • Key catalysts with dates
    • Planned hold period
  4. Kill Criteria step (after bear case):
    • Ask: "What are the core moats, growth drivers, and key threats?"
    • For each dimension the user provides, generate a specific kill criteria condition
    • User reviews and approves/edits the conditions
    • Aim for 4-7 kill criteria total, mix of quantitative and qualitative
    • Also ask for 2-3 key peers (ticker + relationship)
  5. Create thesis.md from template (include NLM attribution fields in frontmatter)
  6. Create thesis.yaml with kill_criteria, peers, supply_chain, quality_grade, idea_source
  7. Pre-fill with portfolio data if position exists

5. Update Subcommand

If /thesis {TICKER} update "{note}":

  • Add entry to Thesis Log table with today's date
  • Confirm update

Thesis Template

# Investment Thesis: {TICKER}

**Company:** {Full company name}
**Last Updated:** {YYYY-MM-DD}

---

## Classification

| Field | Value |
|-------|-------|
| Industry | {Industry} |
| Strategy | {Value/Growth/GARP/Event-Driven/Momentum} |
| Driver | {Valuation/Growth/Momentum/Event/Macro/Technical/Catalyst} |
| Info Source | {Self-Research/Sell-Side/Social Media/News/Podcast/13F/Friend/Earnings/Weekly Meeting/Other} |
| Idea Source | {substack/x/podcast/13f/supply-chain/weekly-meeting/chatgpt/self-research/other} |
| Source Detail | {e.g. "投资观点周报 2025, 2025-12-15 — 首次讨论 ZYN growth"} |
| Position | {LONG/SHORT/NONE} |
| Planned Hold | {30/60/90/180/365} days |

---

## Core Thesis

{2-3 sentences on why you own this}

---

## Bull Case

**Target:** ${TARGET} (+{PCT}% from entry)

1. {Reason 1}
2. {Reason 2}
3. {Reason 3}

## Bear Case

**Stop:** ${STOP} (-{PCT}% from entry)

1. {Risk 1}
2. {Risk 2}
3. {Risk 3}

---

## Key Catalysts

| Catalyst | Expected Date | Status |
|----------|---------------|--------|
| {Event 1} | {YYYY-MM-DD} | Pending |
| {Event 2} | {YYYY-MM-DD} | Pending |
| {Event 3} | {YYYY-MM-DD} | Pending |

---

## Position History

| Date | Action | Qty | Price | Notes |
|------|--------|-----|-------|-------|
| {YYYY-MM-DD} | {BUY/SELL} | {Qty} | ${Price} | {Notes} |

**Current Position:** {X} shares @ ${avg_cost} avg ({pct}% of NAV)

---

## Thesis Log

| Date | Update |
|------|--------|
| {YYYY-MM-DD} | Thesis created |

---

## Next Review

**Date:** {30 days from now}
**Trigger:** Next earnings / Key catalyst date

Example Usage

Create new thesis:

User: /thesis TSM
Claude: No thesis found for TSM. Let me create one.
        Current position: 8,800 shares @ $348.18 avg (15.3% of NAV)

        I'll need a few details:
        1. Industry? [Technology]
        2. Strategy? [Growth/GARP/Value/Event-Driven/Momentum]
        3. Primary driver? [Growth/Valuation/Catalyst/...]
        4. Info source? [Self-Research/Sell-Side/...]
        5. Core thesis?
        6. Bull case + target?
        7. Bear case + stop?
        8. Key catalysts?
        9. Planned hold period?

View existing thesis:

User: /thesis NVDA
Claude: [Displays thesis.md content]

        📊 Current Position Status:
        • Shares: 3,800
        • Avg Cost: $187.97
        • Current: $186.47 (-0.80%)
        • Unrealized: -$5,683
        • % of NAV: 3.2%

        Would you like to update the thesis?

Quick update:

User: /thesis NVDA update "Blackwell shipping ahead of schedule"
Claude: ✓ Added to NVDA thesis log:
        2026-01-27 | Blackwell shipping ahead of schedule

6. Check Subcommand

Syntax: /thesis TICKER check

Review each kill criteria against the latest available information and update check_result + check_note in thesis.yaml.

Workflow:

  1. Read research/companies/{TICKER}/thesis.yaml — extract kill_criteria list
  2. For each criterion:
    • Quantitative: Search for latest data (web search, earnings analysis in Obsidian, NotebookLM history). Present the data and propose pass/warning/fail.
    • Qualitative: Present the condition and last check_note. Ask user for their current assessment (pass/warning/fail) and a brief note.
  3. Update thesis.yaml with new check_result, check_note, and last_checked date
  4. Display summary table with any changes highlighted

Output format:

Kill Criteria Check: PM (2026-02-07)

1. ZYN US share <50% OR category <10% OR competitor +5pp
   Data: Share ~64% (Q4), category +30%, no single competitor threat
   Result: Pass (unchanged)

2. FDA lowers PMTA barriers
   Last note: "PMTA process remains rigorous"
   Your assessment? [pass/warning/fail + note]

   ...

Summary: 5 checked | 4 pass | 1 warning | 0 fail
Updated thesis.yaml

Example:

User: /thesis PM check
Claude: Checking 5 kill criteria for PM...

        1/5: ZYN US share <50% OR category <10% OR competitor +5pp [QT]
        Latest: ZYN share ~64% (Q4 2025), category +30% YoY
        -> Pass (no change)

        2/5: FDA lowers PMTA barriers [QL]
        Last check (2026-02-07): "PMTA process remains rigorous"
        Any update? [pass]

        ...

        Summary: 5/5 pass, 0 warnings, 0 fails
        thesis.yaml updated with new last_checked dates.

7. Challenge Subcommand

Syntax: /thesis TICKER challenge

Stress-test your thesis by surfacing contradicting evidence from NotebookLM history and the Obsidian vault — searching cached Q&A first before making live API calls.

Workflow:

  1. Read existing thesis from research/companies/{TICKER}/thesis.md — extract core thesis, bull/bear cases
  2. Search NLM answer history (free, no daily limit cost):
    cd ~/.claude/skills/notebooklm && python scripts/run.py history_manager.py search --query "{TICKER}"
    Also search company name and common aliases (e.g., NVDA → NVIDIA → 英伟达)
  3. Grep Obsidian vault 导航/NotebookLM/ folder for ticker mentions:
    Grep ~/Documents/Obsidian Vault/导航/NotebookLM/ for TICKER, company name, aliases
  4. Search 13F institutional holdings for smart money signals:
    cd ~/.claude/skills && python shared/13f_query.py {TICKER}
    • Flag any top holders that significantly reduced positions (>20% decrease)
    • Flag if a well-known contrarian fund took a new position
  5. Present "Devil's Advocate" findings: contradicting evidence, risks not in bear case, outdated assumptions
  6. Offer optional deep challenge via NotebookLM:
    cd ~/.claude/skills/notebooklm && python scripts/thesis_challenger.py {TICKER}
    This uploads the thesis to NLM and asks 3 devil's advocate questions with citations. Output saved to Obsidian Vault/导航/NotebookLM/{date}_{TICKER}_challenge.md
  7. Optionally append challenge summary to thesis log via /thesis TICKER update "..."
  8. Suggest updating kill criteria if new risks were discovered

Output format:

## Thesis Challenge: {TICKER}

### Your Current Thesis
> {core thesis from thesis.md}

### Contradicting Evidence from NotebookLM History
| Source | Date | Finding |
|--------|------|---------|
| {notebook_name} | {date} | {relevant excerpt} |

### Risks Not in Current Bear Case
1. {risk from NLM answers not captured in bear case}

### Recommendation
{Keep / Update bear case / Flag for full review}

Example:

User: /thesis NVDA challenge
Claude: Reading NVDA thesis... Core thesis: "AI compute demand sustains pricing power"

        Searching NotebookLM history for NVDA, NVIDIA, 英伟达...
        Found 3 relevant Q&A pairs across 2 notebooks.

        ## Thesis Challenge: NVDA

        ### Your Current Thesis
        > AI compute demand sustains pricing power through 2027

        ### Contradicting Evidence from NotebookLM History
        | Source | Date | Finding |
        |--------|------|---------|
        | AI Infrastructure Economics | 2026-01-25 | Custom ASIC adoption at hyperscalers accelerating |
        | Semiconductor Supply Chain | 2026-02-01 | TSMC capacity expansion may reduce GPU scarcity premium |

        ### Risks Not in Current Bear Case
        1. Google TPU v6 performance parity mentioned in NLM source analysis

        ### Recommendation
        Update bear case — add custom silicon competitive risk

        💡 Want me to query NotebookLM live for deeper analysis?

"Why I Passed" Workflow

Purpose

Track companies you researched but decided NOT to invest in. This is critical for:

  • Auditing your filtering intuition (did passed companies underperform?)
  • Identifying systematic missed opportunities (are you too conservative on a category?)
  • Attribution analysis in /review (what % of researched ideas become positions?)

6. Passed Subcommand

Syntax: /thesis {TICKER} passed

Goal: ≤2 minutes to complete. Low friction is more important than completeness.

Workflow:

  1. Create directory if needed: ~/PORTFOLIO\research\companies\{TICKER}\
  2. Check if passed.md already exists
  3. If NEW: Ask 3 quick questions using AskUserQuestion tool:
    • Q1: "Where did you first see this idea?" Options: Self-Research / Sell-Side / Social Media / Podcast / 13F / Earnings / Friend / Other
    • Q2: "Why are you passing? (1-3 sentences)" Free text — this is the most important field
    • Q3: "Under what conditions would you revisit?" Free text — e.g. "Stock drops >20%" or "New product launch"
  4. Auto-fetch current stock price via Yahoo Finance:
    import yfinance as yf
    price = yf.Ticker(TICKER).info.get('regularMarketPrice', 'N/A')
  5. Create passed.md from template below
  6. Confirm creation with price recorded

If UPDATE (/thesis TICKER passed update): append a new entry to the Passed Log table.

Passed Template

---
status: passed
ticker: {TICKER}
first_seen: {YYYY-MM-DD}
source: {info_source from Q1}
price_at_pass: {current_price}
tags: [passed, {sector}]
---

# Passed: {TICKER}

**Company:** {Full company name — look up via yfinance}
**Date:** {YYYY-MM-DD}
**Price at Pass:** ${price_at_pass}

---

## Source

| Field | Value |
|-------|-------|
| Info Source | {Q1 answer} |
| Source Detail | {optional detail from Q1} |

---

## Why I Passed

{Q2 answer — 1-3 sentences}

---

## Revisit Trigger

{Q3 answer — conditions under which to reconsider}

---

## Passed Log

| Date | Price | Note |
|------|-------|------|
| {YYYY-MM-DD} | ${price} | Initial pass — {brief reason} |

---

## Price Check (auto-updated by /review)

_This section will be populated by the monthly passed review._

Example: Passed

User: /thesis TSLA passed

Claude: Recording a pass on TSLA.

        Q1: Where did you first see this idea?
        > [Social Media]

        Q2: Why are you passing?
        > Valuation too stretched at 80x forward. FSD revenue model
        > still unproven. Elon distraction risk from other ventures.

        Q3: Under what conditions would you revisit?
        > Stock drops >30% or FSD achieves Level 4 regulatory approval.

        Current price: $342.50

        Created: PORTFOLIO/research/companies/TSLA/passed.md
        Price recorded: $342.50 on 2026-02-07

Example: Passed Update

User: /thesis TSLA passed update

Claude: Updating TSLA passed record.
        Current price: $285.00 (was $342.50 at initial pass, -16.8%)

        What's the update?
        > Dropped significantly but FSD still no Level 4. Watching.

        Added to Passed Log:
        2026-04-15 | $285.00 | Dropped 16.8% but FSD still no Level 4. Watching.

Key Files

  • research/companies/{TICKER}/thesis.md - Active thesis document (markdown, detailed)
  • research/companies/{TICKER}/thesis.yaml - Structured data (kill criteria, peers, conviction, sizing)
  • research/companies/{TICKER}/passed.md - Why I Passed record
  • research/companies/{TICKER}/notes/ - Additional research notes
  • decisions/trades/*_{TICKER}.md - Related trade logs

thesis.yaml Schema

ticker: TICKER
invalidation_price: 0         # Price-based stop (read by BiasEngine)
invalidation_window_days: 2
bear_case_1: "..."
bear_case_2: "..."
bull_case: "..."
horizon_days: 365
conviction: 4                  # 1-5 scale
target_weight_pct: 8
created_date: YYYY-MM-DD

# --- Phase 4: Idea Attribution (NLM-assisted) ---
idea_source: weekly-meeting    # substack/x/podcast/13f/supply-chain/weekly-meeting/chatgpt/self-research/other
source_detail: "投资观点周报 2025, 2025-12-15"
source_link: "[[2025-12-15 - 投资观点周报]]"
nlm_citation: "首次在 Week 50 讨论,初始情绪看多..."  # NLM verbatim citation
first_seen: 2025-12-15         # Date idea first appeared
first_position: 2026-02-01     # Date first trade was made (auto-filled by /trade)

kill_criteria:
  - condition: "Description of what would invalidate thesis"
    type: quantitative         # or qualitative
    threshold: "<50%"          # quantitative only
    status: active             # active / triggered / retired
    last_checked: YYYY-MM-DD
    check_result: pass         # pass / warning / fail / unchecked
    check_note: "Brief rationale for the assessment"

peers:
  - ticker: "BATS"
    relationship: "Global competitor"

supply_chain:
  - "Key regulatory/supply dependency"

quality_grade: "B"             # A/B/C based on KC clarity + data availability

Framework Coverage Integration

In Thesis Template

Add a ## Framework Coverage section at the end of the thesis template, after Thesis Log:

## 📐 Framework Coverage

| # | Section | Level | Key Source |
|---|---------|-------|-----------|
| S1 | 📈 Market & Growth | ✅ | earnings Q4, CAGNY |
| S2 | 🏟️ Competitive Landscape | ✅ | supply chain, earnings |
| ... | ... | ... | ... |

Score: 83% (7/9 covered) | Last scan: YYYY-MM-DD
Gaps: S5 管理层

In thesis.yaml Schema

Add framework_coverage field:

framework_coverage:
  score: 83
  last_scanned: 2026-02-07
  gaps: [S5]
  covered: [S1, S2, S3, S4, S6, S7, S8]

Auto-Scan on Create

After creating a new thesis (step 4B), automatically run:

cd ~/.claude/skills && python shared/framework_coverage.py scan TICKER --brief

Show the coverage result as a gap alert:

"📐 Framework coverage: 45% — S3 护城河, S5 管理层, S6 估值, S7 风险 需要更多研究。 运行 /research TICKER --deep 可针对性填补盲区。"

In /thesis TICKER check

Add framework coverage check after kill criteria check (step 4 in check workflow):

cd ~/.claude/skills && python shared/framework_coverage.py scan TICKER --format json

If any section has degraded since last scan → flag it. If framework_coverage.last_scanned is >30 days ago → suggest re-scan.

Integration with /trade

When a trade is logged via /trade, the thesis position history should be updated automatically if the thesis exists.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.