How do I install the Smart Money Flow skill?

Run npx skillmds@latest add kuntal-r-d/smart-money-flow in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Smart Money Flow safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Smart Money Flow?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Smart Money Flow free to use?

Yes. Installing skills from SkillMD is free. This skill is licensed under Apache-2.

Who published Smart Money Flow?

kuntal-r-d (@kuntal-r-d) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Smart Money Flow by kuntal-r-d — Agent Skill

Name: Smart Money Flow
Author: kuntal-r-d

Smart Money Flow

Tracks institutional, foreign, sponsor/director, and fund-manager moves for a DSE stock from public shareholding disclosures and returns an accumulation/distribution score, rating, and reasoning. Use when the user asks about smart money, institutional buying/selling, foreign inflow/outflow, sponsor or director (insider) activity, shareholding changes, which funds hold a stock, or whether big players are accumulating a Dhaka Stock Exchange ticker. Public data only.

kuntal-r-d Updated Sep 17, 2026

File contents

Smart-Money Flow

Prompt-first skill. Read public disclosures by the rules below. scripts/analyze.py is OPTIONAL. Hard guardrail: PUBLIC disclosure data only — never infer or fabricate a private broker book. If data is missing, say so and lower confidence.

Role & objective

You assess accumulation vs distribution from public shareholding-pattern changes and disclosed fund moves, returning a score (−1..+1), confidence and rating.

When to use

"Are institutions accumulating GP?", "any foreign inflow?", "sponsor/insider selling?", "which funds hold this?". Feed the score into signal-synthesizer.

Inputs you need

shareholding — array of monthly snapshots with month and % for sponsor, govt, institution, foreign, public. Need ≥2 months for a trend.
(optional) funds — {name, ticker_weight_pct, prev_weight_pct, track_record_3y, manager}.

If no shareholding: return score 0, confidence 0.1, rating neutral, flag no_disclosure_data.

Method — month-over-month deltas (significance = ±2.0 pp)

Compare the latest two months. For each holder type:

Institution: ≥ +2pp → +0.3 (accumulation); ≤ −2pp → −0.3 (distribution).
Foreign: ≥ +2pp → +0.3 (inflow); ≤ −2pp → −0.3 (outflow); else small +0.1 / −0.1 by sign.
Sponsor/director: ≤ −2pp → −0.4 + flag sponsor_director_selling (RED FLAG); small cut −0.15; ≥ +2pp → +0.25 (insider buying); small rise +0.1.
Funds (optional): weight raised → up to +0.3 scaled by 3-year track record; weight cut → −0.1.

Clamp the total to [−1, +1].

Scoring rubric

Rating: score ≥ 0.2 → accumulation · ≤ −0.2 → distribution · else neutral.

Confidence starts 0.6, capped at 0.70 (monthly cadence is stale → always flag monthly_disclosure_lag); −0.25 if only one month (single_month_no_trend); −0.05 if no fund data.

Output (emit this Thinking Card)

{ "skill": "smart-money-flow", "ticker": "..", "mode": "investment", "as_of": "..",
  "score": 0.0, "confidence": 0.0, "rating": "accumulation|distribution|neutral",
  "key_metrics": { "latest_deltas": {}, "latest_month": "..", "funds_featuring": [],
    "significance_threshold_pp": 2.0 },
  "reasoning": ["dated MoM moves"], "flags": ["monthly_disclosure_lag", "..."],
  "disclaimer": "Educational analysis only. Not financial advice." }

DSE pitfalls

Monthly lag: DSE shareholding is disclosed monthly — it's stale by construction; never present it as real-time and keep confidence ≤ 0.70.
Sponsor selling is the strongest negative — insiders know most; weight it heavily.
No private data: if foreign/institution fields are absent, score them as unknown, not zero-bad.

Optional precision helper

python3 scripts/analyze.py --input data.json --pretty

Returns the per-holder deltas, fund featuring, score/confidence/rating.

Worked example

Latest month: institution +3pp, foreign +2.5pp, sponsor flat → +0.3 +0.3 = +0.6 → accumulation, confidence 0.6 (capped), flag monthly_disclosure_lag.

References

Methodology & guardrails: references/METHODOLOGY.md. Output is educational analysis only, never financial advice.

---name: smart-money-flowdescription: Tracks institutional, foreign, sponsor/director, and fund-manager moves for a DSE stock from public shareholding disclosures and returns an accumulation/distribution score, rating, and reasoning. Use when the user asks about smart money, institutional buying/selling, foreign inflow/outflow, sponsor or director (insider) activity, shareholding changes, which funds hold a stock, or whether big players are accumulating a Dhaka Stock Exchange ticker. Public data only.license: Apache-2.0---# Smart-Money Flow> **Prompt-first skill.** Read public disclosures by the rules below. `scripts/analyze.py`> is OPTIONAL. **Hard guardrail: PUBLIC disclosure data only — never infer or fabricate a> private broker book.** If data is missing, say so and lower confidence.## Role & objectiveYou assess accumulation vs distribution from public shareholding-pattern changes anddisclosed fund moves, returning a score (−1..+1), confidence and rating.## When to use"Are institutions accumulating GP?", "any foreign inflow?", "sponsor/insider selling?","which funds hold this?". Feed the score into `signal-synthesizer`.## Inputs you need- **`shareholding`** — array of monthly snapshots with `month` and `%` for `sponsor`, `govt`, `institution`, `foreign`, `public`. Need ≥2 months for a trend.- *(optional)* **`funds`** — `{name, ticker_weight_pct, prev_weight_pct, track_record_3y, manager}`.If no `shareholding`: return score 0, confidence 0.1, rating neutral, flag `no_disclosure_data`.## Method — month-over-month deltas (significance = ±2.0 pp)Compare the latest two months. For each holder type:- **Institution:** ≥ +2pp → +0.3 (accumulation); ≤ −2pp → −0.3 (distribution).- **Foreign:** ≥ +2pp → +0.3 (inflow); ≤ −2pp → −0.3 (outflow); else small +0.1 / −0.1 by sign.- **Sponsor/director:** ≤ −2pp → **−0.4 + flag `sponsor_director_selling` (RED FLAG)**; small cut −0.15; ≥ +2pp → +0.25 (insider buying); small rise +0.1.- **Funds (optional):** weight raised → up to +0.3 scaled by 3-year track record; weight cut → −0.1.Clamp the total to [−1, +1].## Scoring rubricRating: score ≥ 0.2 → **accumulation** · ≤ −0.2 → **distribution** · else **neutral**.**Confidence** starts 0.6, **capped at 0.70** (monthly cadence is stale → always flag`monthly_disclosure_lag`); −0.25 if only one month (`single_month_no_trend`); −0.05 if no fund data.## Output (emit this Thinking Card)```json{ "skill": "smart-money-flow", "ticker": "..", "mode": "investment", "as_of": "..", "score": 0.0, "confidence": 0.0, "rating": "accumulation|distribution|neutral", "key_metrics": { "latest_deltas": {}, "latest_month": "..", "funds_featuring": [], "significance_threshold_pp": 2.0 }, "reasoning": ["dated MoM moves"], "flags": ["monthly_disclosure_lag", "..."], "disclaimer": "Educational analysis only. Not financial advice." }```## DSE pitfalls- **Monthly lag:** DSE shareholding is disclosed monthly — it's stale by construction; never present it as real-time and keep confidence ≤ 0.70.- **Sponsor selling is the strongest negative** — insiders know most; weight it heavily.- **No private data:** if foreign/institution fields are absent, score them as unknown, not zero-bad.## Optional precision helper```bashpython3 scripts/analyze.py --input data.json --pretty```Returns the per-holder deltas, fund featuring, score/confidence/rating.## Worked exampleLatest month: institution +3pp, foreign +2.5pp, sponsor flat → +0.3 +0.3 = +0.6 → **accumulation**,confidence 0.6 (capped), flag `monthly_disclosure_lag`.## ReferencesMethodology & guardrails: [references/METHODOLOGY.md](references/METHODOLOGY.md).Output is educational analysis only, never financial advice.

Frequently asked questions