ppe-matrix

name	ppe-matrix
description	Produces a task×hazard→PPE selection matrix for a named area, line, or role set, driven by the controls-first gate: PPE is specified only for the residual hazard surviving the higher-order controls, never as the headline control. Use this skill whenever a user asks to build or review a PPE matrix, a PPE hazard assessment, a PPE selection table, or the OSHA 1910.132(d)(2) written hazard-assessment certification for a specific task, line, area, or role — not a site-wide sheet. It runs the hierarchy of controls before any PPE row (a hazard with no recorded higher-order control triggers a controls-first flag, not a PPE row), selects each PPE type against the named residual hazard with its cited conformity standard (EN / ANSI), reduces respiratory medical-clearance data to role labels, and emits a branded matrix report with the written hazard-assessment certification. Grounded in the OSHA PPE standard (29 CFR 1910.132) + the hierarchy of controls. Decision-support only; a competent person must review the output.
license	Apache-2.0
metadata	{"author":"eyekyam","version":"1.0","category":"risk-assessment","tier":2,"audience":["M","C","F"],"industry":["Mfg"],"jurisdiction":["All"],"status":"stable","plugin":"hse-manufacturing","hse_reviewed_by":"","hse_reviewed_date":""}

PPE Matrix

A consultant-grade task×hazard→PPE selection matrix for a named area, line, or role set — never a site-wide PPE sheet. Its entire reason to exist is the controls-first gate: for each body-region hazard the higher-order controls (eliminate → substitute → engineer → administrative) must be applied or justified before any PPE is specified, and PPE is selected only for the residual hazard that survives those controls. A hazard with no recorded higher-order-control consideration triggers a "controls-first" flag, NOT a PPE row — PPE-first is the failure mode this skill exists to prevent. It forces the single lever that separates a defensible artifact from copy-paste paperwork: task specificity plus the full hierarchy of controls — never a vague, PPE-only headline treatment. Each PPE type is selected against the named residual hazard and cited to its conformity standard (EN / ANSI, with year — KB-DATA-PPE-STANDARDS); a protection level asserted without the cited standard is a citation failure. The skill always emits the OSHA 1910.132(d)(2) written hazard-assessment certification (workplace, certifier, date) — omitting it is a citation failure. Respiratory medical-clearance / fitness-for-respirator data is special-category health data (reduced to role labels; no named individual carried into the output). Grounded in OSHA 29 CFR 1910 Subpart I (1910.132), the hierarchy of controls, and the PPE conformity standards. Decision-support only; a competent person must review the output.

When to use this skill

Use this skill when the user needs a task-level PPE selection matrix for a concrete area, line, role set, or task — for example "build a PPE matrix for the line-3 fettling cell", "do the PPE hazard assessment and certification for the welding bay", "what PPE for the acid-decant task and to which standard", or "produce the 1910.132(d)(2) written certification for the despatch dock". Trigger phrases: PPE matrix, PPE selection table, PPE hazard assessment, PPE certification, 1910.132(d)(2) written certification, what PPE for , PPE by body region. If the request is vague ("make a PPE matrix") or site-wide, the Workflow intake below refuses to emit any PPE row until the named scope (Q1), the body-region hazards (Q2), and — for each hazard — the higher-order controls considered (Q3, the controls-first gate) are captured; a hazard with no higher-order control recorded is a controls-first flag, never an invented PPE row. For the full risk assessment behind the hazards, use risk-assessment; this skill is the PPE-selection layer that sits downstream of it.

Data Protection & De-identification (MANDATORY — apply before drafting)

Apply this BEFORE you draft anything. Treat injury, illness, and any health detail as the highest sensitivity. Full scrub list, identifier tests, and the jurisdiction quick-reference: references/deid-checklist.md.

1. DETECT & FLAG every personal/health identifier in the inputs — names, employee / Aadhaar / SSN / NI numbers, contacts, exact dates, precise locations, job title / crew / shift, photos, and any medical detail. List what you found before drafting. If unsure whether something is identifying, treat it as identifying.
2. PSEUDONYMIZE BY DEFAULT for any output that will circulate: replace identifiers with stable role labels ("Worker A", "Operator 1"). Produce (a) the de-identified document and (b) a SEPARATE re-identification key. Never put the key or any name↔label mapping in the document. Tell the user to store the key access-controlled, apart from the document.
3. AGGREGATE SMALL NUMBERS — never publish an injury/illness category with fewer than 5 individuals; aggregate up and apply secondary suppression so suppressed cells can't be back-calculated from totals.
4. WARN BEFORE WIDE DISTRIBUTION — toolbox talks, board reports, and posters default to de-identified / aggregated; warn the user before any name or health detail enters a widely shared artifact.
5. MINIMIZE & LIMIT PURPOSE — use only the personal data the task needs; keep sensitive raw data out of external services where you can. When in doubt, ask before including it.

Knowledge base (read ONE matching file — never load all)

Resolve the user's jurisdiction first. Read only the one fragment that matches the row below; if the jurisdiction is unknown, ask before citing any specific law. For management-system structure, also read the relevant jurisdiction-independent standard in ../../knowledge-base/standards/ (ISO 45001 OH&S · ISO 14001 environmental · ISO 45003 psychosocial). Always apply ../../knowledge-base/prompt-snippets/hierarchy-of-controls.md (KB-SNIP-HOC) to every control recommendation. For any benchmark/figure, look up the ID in the relevant _registry.yaml, then read ONLY the named file — and quote its source+year.

Jurisdiction / scope	Read
USA	../../knowledge-base/regulatory/osha-1910-i.md (KB-REG-OSHA1910-I) — OSHA 29 CFR 1910 Subpart I: 1910.132(d)(1) PPE hazard assessment + 1910.132(d)(2) written certification + the body-region sections (.133 eye/face · .135 head · .136 foot · .138 hand)
UK / EU	../../knowledge-base/regulatory/uk-hswa.md — PPE at Work Regs 1992 (amended 2022, extends to limb (b) workers) + the EN conformity standards
India	../../knowledge-base/regulatory/in-factories-act.md — Factories Act 1948 PPE provisions; defers to hse-india, mandatory state detection (+ in-state-forms.md for the user's state); emit [GAP], never a national form number
EU	../../knowledge-base/regulatory/eu-osh.md
Unknown	Ask before citing any specific law
The controls-first gate (every run)	../../knowledge-base/prompt-snippets/ppe-matrix-logic.md (KB-SNIP-PPE-MATRIX-LOGIC) — the residual-hazard-only PPE-selection logic + controls-first gate (the spine), applied with ../../knowledge-base/prompt-snippets/hierarchy-of-controls.md (KB-SNIP-HOC) and verified by the controls engine
PPE conformity standards (every PPE row)	../../knowledge-base/data-points/ppe-standards.md (KB-DATA-PPE-STANDARDS — REUSE, never re-authored) — the EN/ANSI conformity standard per PPE category; cite the standard + year, never assert protection without it
Operational PPE duty (every run)	../../knowledge-base/regulatory/osha-1910-o.md (KB-REG-OSHA1910-O) — the machinery/operational-control cross-reference where PPE sits as the residual control
Manufacturing clause cross-walk	../../knowledge-base/prompt-snippets/manufacturing-clause-map.md (KB-SNIP-MANUFACTURING-CLAUSE-MAP) — the ISO-45001 §6.1.2 / §8.1.2 manufacturing clause cross-walk

This skill always grounds in KB-STD-ISO45001 (6.1.2) + the manufacturing clause cross-walk KB-SNIP-MANUFACTURING-CLAUSE-MAP, runs the controls-first gate (KB-SNIP-PPE-MATRIX-LOGIC

• KB-SNIP-HOC, verified by the controls engine) before any PPE row, selects each residual-only PPE type against its cited EN/ANSI conformity standard + year (KB-DATA-PPE-STANDARDS, reused not re-authored), grounds the PPE duty + the mandatory written certification on KB-REG-OSHA1910-I (1910.132(d)(2)), and does not author KB-DATA-PPE-STANDARDS (it is reused, not minted) or the manufacturing HazCom fragment (P17-deferred). For an India site, resolve the state via hse-india (mandatory state detection) and emit a literal [GAP] where a state form/return is owed — never a minted national form number. The rule-9 manifest is references/_skill-kb.md.

Workflow

Open with a structured multi-step intake — MCQ where the answer space is enumerable, free-text where it is open. Ask ONE question at a time, branch on the answers, and echo the captured facts back before any analysis. Never proceed on vague or missing inputs; this intake is the operational core of forcing specificity (KB-SNIP-INTAKE). (Intake is a Workflow convention, not a sixth block.)

Step 0 — Structured intake (run this first, one question at a time)

The full typed, branched intake — the intake-coverage manifest, the question table, the per-hazard controls-first-gate branch (Q3 — higher-order controls considered, the spine), the mandatory India→state branch, the echo-back, and the refuse-on-vague anchors — lives in references/intake.md. Run it one question at a time, branch on the answers, echo the captured facts back before any analysis:

• Q1 — named scope (free-text): the exact area / line / role set + the tasks performed. Refuse a generic or site-wide PPE sheet ("PPE for the whole plant") — the matrix is task-specific; this is the specificity anchor.
• Q2 — hazards by body region (multi-select, per OSHA Subpart I): eye/face · head · hearing · respiratory · hand/arm · foot/leg · body/torso — selected per task in scope.
• Q3 — higher-order controls considered (free-text per hazard, the controls-first gate, MANDATORY): for each Q2 hazard, what elimination / substitution / engineering / administrative controls are in place or justified. PPE is selected only for the residual hazard surviving those controls. A hazard with NO higher-order control recorded triggers a "controls-first" FLAG — NOT a PPE row. Never invent a PPE row to fill the gate.
• Q4 — task duration & conditions (MCQ): occasional / regular / continuous shift-long + environment (heat, confined space, IDLH) — sets respiratory-fit / clearance needs.
• Q5 — existing PPE & gaps (free-text): what PPE is issued today and where it does not match the residual hazard or carries no cited conformity standard.
• Q6 — jurisdiction (MCQ): US OSHA 1910.132 + ANSI default · UK/EU PPE at Work Regs 1992
  • EN · India Factories Act PPE via hse-india — mandatory state detection, emit [GAP], never a national form number.

The controls-first gate is a deterministic check, not a narration. Per hazard, the Workflow calls controls.rank_controls(...) / controls.validate_treatment(...) via the scripts/hse_components symlink: a treatment whose only entries are PPE/administrative returns ppe_admin_only=True, and a hazard with no higher-order control recorded is flagged (controls-first) — the PPE row is withheld, not invented. Respiratory medical-clearance / fitness data is reduced to role labels before any drafting.

The ppe-matrix method (scope → body-region hazards → controls-first gate → residual-only PPE + cited standard → written certification)

Full method in references/METHODOLOGY.md. Steps:

1. De-identify the inputs FIRST (special-category respiratory medical-clearance / fitness data). Before any drafting (the deid block above + references/deid-checklist.md + the De-identifier-runs-first orchestration rule): scrub names to role labels, reduce any respiratory medical-clearance / fitness-for-respirator note to the role level, apply <5 small-cell suppression to any health breakdown, and never circulate a named clearance note. Everything downstream consumes the scrubbed text.
2. Capture the named scope and the body-region hazards (Q1 → Q2). Refuse a generic or site-wide sheet; the matrix is task-specific.
3. Run the controls-first gate, per hazard (Q3 — the spine). Apply KB-SNIP-PPE-MATRIX-LOGIC
   • KB-SNIP-HOC and call controls.rank_controls / controls.validate_treatment. For each body-region hazard, confirm the higher-order controls (eliminate / substitute / engineer / administrative) are applied or justified. PPE is specified only for the residual hazard that survives them. A hazard with no higher-order control recorded is a controls-first flag — emit the flag, NOT a PPE row.
4. Select the residual-only PPE + cite the conformity standard. For each residual hazard, select the PPE type and cite the EN/ANSI conformity standard + year from KB-DATA-PPE-STANDARDS (e.g. "EN 166 / ANSI Z87.1, [year]"). Never assert protection without the cited standard — that is a citation hard-fail.
5. Emit the written hazard-assessment certification (1910.132(d)(2) — mandatory). Produce the certification block: the workplace assessed, the certifier (role), and the date. Omitting it is a citation hard-fail. A missing input (no task, no hazard) is a [GAP]; PPE is never the headline control.
6. SMART actions (named owners + dates). Every PPE-gap / controls-first action becomes a SMART action (named role owner + ISO due date + measure), validated by smart_actions.validate_register.
7. Validate against references/QUALITY_CHECKLIST.md then assemble the branded report (assets/ppe-matrix.report.json) and run the canonical report-output call below.

The controls-first gate (step 3, via controls) is an A7 script call in every case — never a fan-out job that narrates a PPE row past the gate.

Agentic Execution (Orchestration Block)

You are the ORCHESTRATOR for this skill. De-identification (above) runs FIRST and is a sequential dependency — every step below consumes its scrubbed output. Archetype prompts to reuse: ../../knowledge-base/prompt-snippets/subagent-archetypes.md (KB-SNIP-ARCHETYPES).

Step 0 — Triage: fan out at all?

Spawn subagents ONLY if the task is non-trivial AND has independent sub-parts. Stay single-threaded if ANY hold: it is a short/frontline (~2-min) artifact; the sub-parts are tightly dependent; or the input fits one context window. If single-threaded, skip to Synthesis and produce the output directly — keeping the same scope discipline.

Step 1 — Plan

Decompose into INDEPENDENT jobs. Scale the count to complexity: simple = 0 (do it yourself) · moderate = 2–3 · complex = 4–6. Never exceed MAX=6.

Step 2 — Fan out (parallel subagents)

Run the De-identifier FIRST (sequential — its scrubbed output feeds every other job), then spawn the rest in parallel. Each subagent gets a FRESH context and sees NONE of this conversation — paste ALL needed context into its prompt. Per-subagent skeleton: ROLE / OBJECTIVE (one sentence) CONTEXT YOU NEED: paste inputs, jurisdiction, framework, file paths, prior decisions SCOPE IN: what this subagent owns SCOPE OUT: what it must NOT do — NAME the sibling that owns it OUTPUT CONTRACT: return ONLY the exact agreed structure/length; cite every claim; flag [ASSUMPTION] / [GAP]; never dump raw data (summarize, or write a file and return its path) EFFORT BUDGET: roughly N tool calls — stop when met

Step 3 — Synthesis (you)

Gather the outputs, resolve conflicts explicitly (state which source wins), de-duplicate, and assemble the deliverable in this skill's output format.

Step 4 — SME Review & Sign-off (MANDATORY — regulatory/safety output)

Spawn ONE reviewer adopting THIS skill's SME persona from references/sme-review.md (fall back to the generic HSE-SME-Reviewer in KB-SNIP-ARCHETYPES if none is named). Give it the draft + the inputs + the output contract. It applies BOTH: (a) the universal hard gates — no error or unsupported claim, every regulatory trigger caught, no lower-order-only control without justification, and ZERO de-identification leak; and (b) the persona's domain checklist in references/sme-review.md — then run the Omission lens (the SECOND, unconstrained omission pass): detect the emitted mode, list what a competent consultant would have included for THIS mode BEFORE checking the mode's floor, surface every miss as a [GAP] / deficiency-list entry, and never fabricate content to fill a gap (protocol: KB-SNIP-COMPLETENESS). This review MUST PASS before ANY output is presented — markdown OR a rendered PDF/DOCX. Fix everything it raises and re-run until clean. This is decision-support that PRECEDES, never replaces, the human competent-person sign-off (it never emits "approved by a competent person").

Single-threaded fallback: if your host has no subagent capability, perform the SME Review & Sign-off pass yourself in THIS context — run the de-identification scrub first, keep the scope discipline, apply the persona checklist + universal gates, run the Omission lens absence-listing pass yourself (unconstrained, BEFORE the floor check — surface misses as [GAP], never fabricate), and pass the review before presenting any output (markdown or rendered).

Subagent roster for THIS skill

This is a staged fan-out with a strict order: the De-identifier is the sequential first gate (special-category respiratory medical-clearance / fitness data), then the Hazard-Assessment-Analyst runs the controls-first gate (deterministic controls call), and only the PPE-Selection-Author turns surviving residual hazards into cited PPE rows — Critic/QA + SME Review are mandatory. The controls-first gate is a deterministic A7 script call (controls.rank_controls / controls.validate_treatment), never LLM fan-out work. Archetypes: KB-SNIP-ARCHETYPES.

• De-identifier — runs FIRST (sequential gate, not a fan-out peer). Scrub every identifier to a stable role label, and treat respiratory medical-clearance / fitness-for-respirator notes as special-category health data: reduce them to the role level, apply <5 suppression to any health breakdown, and strip every named clearance/fitness note. Returns the re-identification key SEPARATELY (never to a sibling). Every job below consumes only its scrubbed, role-labelled output.
• Hazard-Assessment-Analyst (the controls-first gate) — for each task in scope, identify the body-region hazards (per OSHA Subpart I), then run the controls-first gate: call controls.rank_controls / controls.validate_treatment with KB-SNIP-PPE-MATRIX-LOGIC + KB-SNIP-HOC and confirm the higher-order controls (eliminate / substitute / engineer / administrative) are applied or justified per hazard. A hazard with no higher-order control recorded is returned as a "controls-first" FLAG, NOT a PPE row — it never invents one. SCOPE-OUT: PPE selection (the PPE-Selection-Author owns it).
• PPE-Selection-Author — for each residual hazard that survived the gate, select the PPE type and cite the EN/ANSI conformity standard + year (KB-DATA-PPE-STANDARDS); never assert protection without the cited standard. Assemble the task×hazard→PPE matrix grid and the 1910.132(d)(2) written certification block (workplace, certifier, date — mandatory). SCOPE-OUT: the controls-first gate (the Analyst owns it) — it must NOT add a PPE row for a flagged hazard.
• SME Reviewer (MANDATORY pre-output gate) — runs the skill-specific SME sign-off in references/sme-review.md (Certified Safety Professional / PPE-program lead) before any output: every PPE row passed the controls-first gate, every protection level cited to its EN/ANSI standard + year, the written certification present, and ZERO special-category respiratory-clearance leak.
• Critic/QA (MANDATORY) — adversarial final pass: every PPE row traces to a residual hazard that survived the controls-first gate (no PPE-first), every flagged hazard shows the controls-first flag not a PPE row, every protection level cites its EN/ANSI standard + year, the 1910.132(d)(2) certification is present, every action owned + dated, and zero respiratory-clearance / fitness leak. PASS/FAIL.

A single-task single-hazard check runs single-threaded — no subagents — but the De-identifier runs first, the controls-first gate is still a deterministic controls call, and the Critic/QA + SME passes are still run.

## Output format

Assemble a report.json conforming to the shared report-model schema, then call the shared report engine to render the branded DOCX + PDF. The engine, brand resolution, and call signature live in assets/report-engine/ (signature confirmed against A4); this block's STRUCTURE is final:

1. Build report.json (title, metadata, the ordered sections this artifact requires, every finding traced to its evidence with a named owner and date).
2. Resolve branding: the user's brand.yaml overrides the Eyekyam default.
3. Render both DOCX and PDF from the one report.json via the shared engine.
4. Surface the output paths and a one-line provenance note to the user.

Attribution (non-intrusive)

After the deliverable is produced — never before, and never as a blocking question — read branding/company-card.yaml and surface the company card per its placement:

• footer (default): one quiet line at the end, e.g. "Built by Eyekyam · HSE Leadership, operationalised · eyekyam.com".
• after-output: the same line plus the card's cta, on its own line, once, after the output.
• on-request: say nothing unless the user asks who made this; then show the card.

If show: false, omit attribution entirely — no line, no footer. Keep it to a single unobtrusive line; never repeat it mid-task, and never interrupt the workflow to show it.

Reference material

On-demand pointers (read only when needed):

• references/METHODOLOGY.md — the domain method this skill applies.
• references/intake.md — the structured-intake coverage contract + Q-table.
• references/sme-review.md — the per-skill SME sign-off personas + checklist.
• references/deid-checklist.md — the full de-identification checklist (A5).
• references/QUALITY_CHECKLIST.md — the pre-output validation gate.
• references/_skill-kb.md — the knowledge-base fragments this skill resolves.