portfolio-eval
Skill0 starsChanged 4 months agoMITView on GitHub
A deeper periodic evaluation of the whole portfolio — per-name weight, unrealized P&L, concentration and sector clustering, portfolio heat/health, and a per-name quality re-check — from a read-only broker MCP (if connected) or a manual-paste snapshot. Renders a beautiful portfolio Artifact where supported. Use for "evaluate my portfolio", "how's my portfolio", periodic portfolio health checks. Broker-agnostic, read-only.
In this repository8 other files it gives its agents.
- autopilotSkill
- checkSkill
- india-dataSkill
- opportunity-screenSkill
- portfolio-watchSkill
- report-artifactSkill
- researchSkill
- watchlist-monitorSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 14
Run scripts/portfolio_eval.py via the capability resolver (Gate MCP → python3 "${CLAUDE_PLUGIN_ROOT}/scripts/kit.py" → JS-paise Artifact → …
- Distinctive: line 24
Never place or suggest placing an order.
- Distinctive: line 25
Never report a holding, price, or metric you did not read/compute this run.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Portfolio Evaluation (health + concentration + quality)
Where portfolio-watch is the daily alert, this is the deeper periodic evaluation: is the portfolio survivable as a whole?
Get holdings — broker-agnostic, fail-closed
Via india-data: a read-only broker MCP (Upstox/Kotak preferred; Zerodha/Groww reads only — the write tools are deny-hooked) → delay_class: live, else a manual-paste snapshot (SYMBOL QTY AVG_COST) → delay_class: manual. If neither is available, say so and STAND DOWN — never guess. Get delayed current prices from Yahoo (labeled delayed). Stamp every reading {value, source, as_of, delay_class}.
Deterministic metrics (do not compute these in prose)
Run scripts/portfolio_eval.py via the capability resolver (Gate MCP → python3 "${CLAUDE_PLUGIN_ROOT}/scripts/kit.py" → JS-paise Artifact → REFUSE) for per-name value, weight, unrealized P&L %, the concentration flag (single-name weight over the 15% cap), total value, and the top name. Portfolio heat/covariance uses the existing heat module.
Procedure
- Composition — per-name weight + unrealized P&L %; flag any single name over the concentration cap; note sector clustering.
- Health — portfolio heat vs the cap; number of names vs
floor(heat/risk); drawdown vs high-water-mark if history is provided. - Per-name quality re-check — run the surveillance/exclusion filter + the fundamental forensic red-flag checks on each holding (from screener /
india-data); flag any that would fail entry today (candidate to trim/exit). - Verdict per name — HOLD / TRIM / EXIT + the single dominant reason.
- Output — where supported, a beautiful portfolio Artifact (weight bars, P&L, concentration/health callouts, per-name verdicts); else a text digest. Append the disclaimer.
Never
- Never place or suggest placing an order. Read-only — the human acts in their broker.
- Never report a holding, price, or metric you did not read/compute this run; never compute the deterministic metrics in prose.
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
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.

