Parallax Halal Screen

Halal / Shariah Screen

When not to use

• General thematic screening → use /parallax-thematic-screen
• Portfolio construction → use /parallax-portfolio-builder
• Non-compliance screening → use /parallax-earnings-quality

Gotchas

• JIT-load _parallax/parallax-conventions.md for RIC resolution, parallel execution, and fallback patterns
• explainmethodology does NOT support shariah/halal — valid concepts are value, quality, momentum, defensive, tactical, overall, factorweighting, scoring
• Shariah thresholds are hardcoded in this skill (commonly used screening ratios) — derive compliance from get_financials data
• getfinancialanalysis (Palepu framework) is async ~2-5 min — warn user before calling
• Financial ratios from get_financials help verify debt/revenue compliance thresholds
• JIT-load _parallax/white-label/integration-pattern.md before the Pre-Render step. Loader call is load_visual_branding() (7-key visual subset; voice structurally excluded — branding["voice"] raises KeyError). Apply §5 (Branding Header) and §7 (About This Report) in Output Format.
• Cost: ~4 tokens single-stock, +5 optional if Palepu is called; ~4-5/holding in portfolio mode. Breakdown in _parallax/token-costs.md.

Screen stocks and portfolios for Shariah compliance using commonly used screening ratios applied to Parallax financial data.

Screening Thresholds

These are the Shariah compliance thresholds applied by this skill. They are commonly used screening ratios, not a claim of conformance to any named published standard. They are applied to data retrieved from get_financials.

| Ratio | Threshold | Pass condition | |-------|-----------|----------------| | Total debt / Total assets | = 33%) and (cash + interest-bearing securities) / total assets (FAIL if >= 33%**).

1. Call get_financials with statement "ratios" to retrieve interest income and revenue figures. Compute: (interest income + non-permissible revenue) / total revenue. FAIL if >= 5%. If >0% but 60% concentration in a single sector but sector_concentration: {} is empty and "well-diversified" is returned, the tool's concentration detection has silently failed — compute concentration client-side from per-holding sectors and flag the tool bug in output. For N60% single-sector share requires ≥5 holdings clustered in one sector, which is unlikely from natural Shariah filtering (which typically excludes ~3 sectors broadly, leaving the residual spread across ≥4–5 sectors); at N≤7 the same share can result from 4 holdings and is consistent with screening outcome, not a tool defect. Tune the cutoff only if production data shows the false-positive rate diverging from this assumption.
2. For non-compliant holdings, call build_stock_universe with "[sector]" to find alternatives, then screen those alternatives through the same compliance check. Do not source alternatives for UNVERIFIED holdings — an unscreened name has not been rejected, so proposing replacements would present a data gap as a compliance finding. Recommend re-running the screen for those names instead.
3. Call get_peer_snapshot on compliant alternatives to score them.

Output Format

• Screening Criteria (commonly used screening ratios as listed above)
• Plain-Language Summary (under audience=client_safe only): 2-3 sentences stating the screen outcome counts in plain terms for a non-specialist reader; factor names carry the §13.3 gloss if used, no cutoff arithmetic, framed per §12 as informational with no directives. State all three counts whenever any name is UNVERIFIED — compliant, non-compliant, and unverified — in plain words (e.g. "2 holdings could not be screened because financial data was unavailable"). A summary reporting only "18 of 20 compliant" while 2 went unscreened is the §4.0 failure this skill exists to prevent, and it is the first line the client reads.
• Compliance Results (table: symbol, status ∈ COMPLIANT / NON-COMPLIANT / UNVERIFIED, reason if non-compliant or unverified). UNVERIFIED is a rendered state, never an omission: give the reason as the field or call that was unavailable (e.g. UNVERIFIED — total revenue unavailable from get_financials(ratios)). Do not collapse the three states into a binary Y/N column, and never leave a screened symbol out of this table.
• Unverified Names (render only if any row is UNVERIFIED): one line per name stating what was missing and that the name is excluded from the Compliant Subset pending a re-screen. Per parallax-conventions.md §4.0 this is a data-integrity warning and is not suppressible under audience=client_safe — a client reading a compliant list must be able to tell "screened and cleared" from "could not be screened."
• Key Ratios (debt/assets %, interest-bearing/assets %, non-permissible revenue %; render each unavailable ratio as UNVERIFIED, not as a blank, a dash, or an estimate)
• Verdict sensitivity: the ratio nearest its cutoff (33% debt/assets, 33% interest-bearing/assets, or 5% non-permissible-revenue) and the arithmetic flip condition, per parallax-conventions.md §11 (internal_analyst mode only per conventions §13.2). Applies only to the three quantitative ratios above — never to the qualitative business-activity screen. Select only among ratios that were actually computed: an UNVERIFIED ratio has no distance to its cutoff, so it can never be the surfaced input, and a name whose overall verdict is UNVERIFIED renders no sensitivity line at all (per §11.2, omitting it is the compliant behavior — never infer a flip number from a missing input).
• Purification Amount (if applicable — percentage of dividends requiring purification)
• Alternatives (for non-compliant holdings: scored compliant replacements in same sector — informational candidates, not replacement instructions)
• Compliant Subset (if portfolio mode: the screened-compliant holdings and their current weights — an informational screen result per conventions §12, not a rebalancing instruction; for construction use /parallax-portfolio-builder then re-screen). Contains verified-compliant names only — UNVERIFIED holdings are excluded per conventions §4.0. If any were excluded, state the count and that weights were renormalised over verified-compliant names, so the subset is never mistaken for a full screen of the portfolio.
• Branding Header (only if white_label_active AND client_name != "") — single line at the very top: **** Shariah screen. Logo handling per integration-pattern.md §5.
• About This Report (always present): one line stating branding state per integration-pattern.md §7. If a logo was skipped, append Logo on file: as a second About This Report line.

Pre-Render — Load white-label branding

Load _parallax/white-label/integration-pattern.md §2 and compute white_label_active + client_name per that section. Apply §5 (Branding Header) and §7 (About This Report) when composing the Output Format.

AI-interaction disclosure (required regardless of view state): Render parallax-conventions.md §9.2 immediately above the disclaimer below.

Render the standard disclaimer verbatim from parallax-conventions.md §9.1.

> These are analytical outputs based on commonly used Shariah screening ratios (debt and interest-bearing-asset thresholds measured against total assets) applied to Parallax financial data, not investment advice or a fatwa. Consult a qualified Shariah advisor for binding rulings.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: bencharoenwong
• Source: bencharoenwong/parallax-workflows
• License: MIT

Install and usage instructions live in the source repository linked above.