Conviction calibration engine

Skill ternary-ai/skills/skills/advanced/conviction-calibration-engine

A collection of agent skills for investment finance

Install

npx -y skills add ternary-ai/skills --skill conviction-calibration-engine

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Aligns position size with current thesis conviction, edge, and asymmetry; adjusts only when evidence changes. Use during ongoing portfolio management or after a material thesis update.

SKILL.md

3.4 KB, as published. Nobody here has run it

Conviction Calibration Engine

Purpose: Continuously align size with edge.

Trigger: Ongoing portfolio management or post-update review.

Inputs:

Expected asymmetry
Thesis updates
Market mispricing changes
Risk profile changes

Data Sources

Follow this order — do not ask the user which source to use:

Context first: Check <stock_context> and <acquired_data> for the current thesis conviction score, upside/downside targets, and current position size vs portfolio. If asymmetry data and current thesis are present → proceed directly to Process.
Fetch if missing: Call:
- GLOBAL_QUOTE — current price to recompute upside/downside %
- OVERVIEW — updated analyst target for consensus anchor
Ask only if both fail: Call request_user_input asking for the specific missing field only (e.g. "Please provide your current upside and downside price targets, and your current position size as % of portfolio").

Process:

Recalculate expected asymmetry.
Adjust size only if thesis improves, mispricing widens, or risk declines.
Prevent creeping overconfidence with explicit criteria.

Output Format

Write a dedicated thesis section in markdown format as follows:

## Conviction Calibration Update
**Analysis Date**: {Date}

### Current Position Status
- **Current Size**: X% of portfolio
- **Current Price**: $X
- **Original Entry**: $X (state date if available)
- **Unrealized P/L**: +/-X% or $Y

### Updated Expected Payoff
- **Upside Target**: $X (+Y% from current)
- **Downside Target**: $X (-Y% from current)
- **Upside/Downside Ratio**: X:1
- **Expected Value**: $X per dollar invested

### Thesis Conviction Assessment
- **Conviction Level**: [Low / Medium / High]
- **Conviction Change**: [Strengthened / Unchanged / Weakened] since last review
- **Edge Status**: [Widened / Stable / Narrowing]
- **Key Evidence Changes**: [bullet list of what changed — earnings, news, management action, competitor move, macro shift]

### Recommended Action
- **Action**: [Hold current size / Add X% / Trim X% / Exit]
- **Target Size**: X%–Y% of portfolio (vs X% current)
- **Rationale**: [2–3 sentences explaining why the size adjustment is warranted — must reference specific thesis updates, asymmetry changes, or risk profile shifts]

### Size Adjustment Criteria Applied
[State which trigger condition was met:]
- ☐ Thesis materially improved → warrant size increase
- ☐ Mispricing widened → warrant size increase  
- ☐ Risk declined → warrant size increase
- ☐ Thesis weakened → warrant size decrease
- ☐ Risk increased → warrant size decrease
- ☐ No material change → hold current size

### Action Summary
"[Hold / Increase / Decrease] position to X% of portfolio. [Brief justification in one sentence.]"

Thesis Field Rule: Always populate the thesis field in the JSON output with the complete markdown section above. This is an advanced skill — thesis upserting is mandatory. ⚠️ Extended thinking is discarded — copy the complete analysis into the thesis field; it is the ONLY output that reaches the Thesis panel.

Output: