Determines and prepares mandatory HSE regulatory returns for a named organisation across jurisdictions — OSHA recordkeeping (Forms 300/300A/301) and electronic submission, UK RIDDOR reports, and EU equivalents. Use this skill whenever a user asks which incidents are recordable or reportable, how to complete OSHA 300/300A/301, whether something is RIDDOR-reportable, what the reporting deadlines are, or to prepare a periodic statutory return. It applies the jurisdiction's recordability/reportability tests, identifies the correct form and deadline, and prepares a de-identified return with the evidence trail. For India, it defers entirely to the hse-india skills (state forms, accident notices) and never hard-codes national form numbers; state detection is mandatory. Decision-support only; a competent person must review the output.
Determines and prepares mandatory HSE regulatory returns for a named organisation across jurisdictions — OSHA recordkeeping (Forms 300/300A/301) and electronic submission, UK RIDDOR reports, and EU equivalents. Use this skill whenever a user asks which incidents are recordable or reportable, how to complete OSHA 300/300A/301, whether something is RIDDOR-reportable, what the reporting deadlines are, or to prepare a periodic statutory return. It applies the jurisdiction's recordability/reportability tests, identifies the correct form and deadline, and prepares a de-identified return with the evidence trail. For India, it defers entirely to the hse-india skills (state forms, accident notices) and never hard-codes national form numbers; state detection is mandatory. Decision-support only; a competent person must review the output.
A consultant-grade HSE skill that determines and prepares a mandatory statutory
return for a named organisation, with a jurisdiction branch — US OSHA 29 CFR
1904 recordkeeping (Forms 300 log / 300A annual summary / 301 incident
report + the 1904.41 electronic submission), UK RIDDOR 2013 (specified injuries,
over-7-day incapacitation, reportable diseases, dangerous occurrences, the
report-within-15-days rule), , and (state forms, accident notices) with . It forces the single lever
that separates a defensible return from guesswork: every recordability/reportability
decision is , tied to the , and prepared as a .
A statutory return holds injured-person identity — so this is one of the
highest-de-identification-sensitivity skills (see ).
## Output format
EU member-state equivalents
India → defers
entirely to the hse-india skills
mandatory
state detection and no hard-coded national form numbers
traced to the named test it applied
correct form + the
exact deadline
de-identified return with an evidence trail
by law
references/deid-checklist.md
When to use this skill
Use this skill when the user needs to determine or prepare a specific statutory HSE
return for a named organisation and a real incident or reporting period. Trigger
scenarios that reinforce the description:
"Is this incident recordable on the OSHA 300 log?" / "Complete our 300A
annual summary." / "Fill the 301 incident report." / "Do we have to submit
300A electronically this year?"
"Is this RIDDOR-reportable?" / "Which RIDDOR category — specified injury,
over-7-day, disease, dangerous occurrence — and what's the deadline?"
For India: "Prepare our state Factories Act return / accident notice" → this
skill runs state detection and routes to hse-india (factories-act-returns /
india-accident-notice / india-state-form-finder); it never mints a national
form number.
If the request is vague, the Workflow intake below refuses to give a
reportable/recordable yes-or-no until the incident facts the test needs are captured,
and never asserts a deadline without the jurisdiction.
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.
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.
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.
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.
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.
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.
Resolve the jurisdiction FIRST (intake Q1), then read only that row. The
recordability/reportability decision logic + the per-jurisdiction deadlines live in
KB-DATA-RECORDABILITY-TESTS (each test cited); the determine→form→deadline→prepare
method is KB-SNIP-RETURNS-METHOD; the operations clause cross-walk is
KB-SNIP-OPS-CLAUSE-MAP. For India, do not read a national form value from
here — run mandatory state detection and route to hse-india (the three-tier
degrade below); an unverified form is a [GAP], never a minted number.
Jurisdiction / scope
Read
USA — OSHA 1904 (300 / 300A / 301 + electronic)
../../knowledge-base/regulatory/us-osha.md (KB-REG-US-OSHA — the 1904 recordkeeping leg)
India — DEFERS to hse-india (state detection MANDATORY)
../../knowledge-base/regulatory/in-factories-act.md + ../../knowledge-base/regulatory/in-state-forms.md (for the detected state) + ../../knowledge-base/regulatory/in-portals.md — route to the hse-india skills factories-act-returns / india-accident-notice / india-state-form-finder; never a national form number; unverified → [GAP]
Unknown
Ask before citing any specific law, form, or deadline
India three-tier deferral (the D-04 graceful degrade — never hard-blocks):
Subagent — where subagents are supported, spawn one that runs the relevant
hse-india skill (factories-act-returns for the state return, india-accident-notice
for the statutory accident notice, india-state-form-finder to resolve the form),
and fold its output into the integrated return.
Main-thread inline — on hosts without subagents, the main thread reads
KB-REG-IN-FACTORIES / KB-REG-IN-STATEFORMS / KB-REG-IN-PORTALS inline and
produces the India leg single-threaded.
Routing prose + [GAP] — if hse-india is not installed at all, emit routing
prose naming the skill to run (factories-act-returns / india-accident-notice /
india-state-form-finder) + the KB pointer + a [GAP] for any unverified form
value. At every tier: mandatory state detection first; never mint a national form
number; never hard-block.
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,
and the refuse-on-vague anchors — lives in references/intake.md. Run it one
question at a time, branch on the answers, and echo the captured facts back before
any analysis. The five questions (per the build spec):
Jurisdiction (MCQ: US OSHA · UK RIDDOR · EU member-state · India → defers to
hse-india, runs mandatory state detection). Branch per jurisdiction's tests.
Return type (MCQ per jurisdiction: US — 300 log / 300A summary / 301 incident
report / electronic 300A; UK — specified injury / over-7-day / disease / dangerous
occurrence).
Incident facts (free-text: nature, outcome, days away/restricted, whether a
specified injury — these drive the recordability/reportability test; the
specificity anchor).
Organisation profile (size / SIC — for OSHA electronic-submission applicability).
Period (for periodic returns / the 300A posting window).
Refuse-on-vague anchors (the core-value lever):
Never give a "reportable / recordable yes-or-no" until the incident facts the test
needs (Q3) are captured — ask again, or record [ASSUMPTION] / [GAP]; never
invent a fact to force a determination.
Never assert a deadline without the jurisdiction (Q1) — the deadline is a
function of the jurisdiction's test (e.g. RIDDOR 15-day rule vs OSHA 300A posting
window); stating one without the other is a defect.
For India, Q1 = India → mandatory state detection then route to hse-india;
never cite a national form number.
The recordability/reportability method (determine → form → deadline → prepare)
Full method in references/METHODOLOGY.md (KB-SNIP-RETURNS-METHOD). Steps:
De-identify the inputs FIRST — before any drafting (the deid block above +
the De-identifier-runs-first orchestration rule). A statutory return holds
injured-person identity by law: the legally-required submission is named and
access-controlled; every internal/shared/working copy is role-labelled, with
<5 suppression on any aggregated summary, and no re-identification key embedded
(the reinforced references/deid-checklist.md).
Apply the jurisdiction's recordability/reportability test — for the resolved
jurisdiction (Q1), run the cited decision logic in KB-DATA-RECORDABILITY-TESTS
against the incident facts (Q3): OSHA 1904 recordability (work-relatedness, new
case, beyond first-aid, days-away/restricted); RIDDOR reportability (specified
injury reg. 4/Sch. 1, over-7-day reg. 4, disease reg. 8/9, dangerous occurrence
reg. 7/Sch. 2). State the verdict with the test cited; flag [GAP] where a
fact is missing — never assume to force a "yes" or a "no".
Identify the correct form — from the verdict + return type (Q2): OSHA 300 / 300A
/ 301 (+ 1904.41 electronic where the org profile (Q4) triggers it); the RIDDOR
category; the EU member-state equivalent. For India, do not name a form here
— route to hse-india (three-tier degrade above); unverified → [GAP].
Calculate the deadline — from the jurisdiction's rule (RIDDOR within 15 days;
OSHA 300A posted Feb 1–Apr 30 and electronically submitted by the 1904.41 date) and
the period (Q5). Never assert a deadline without the jurisdiction.
Prepare the de-identified return + the evidence trail — assemble the return
from the role-labelled facts, each determination linked to the evidence item
that supports it; any follow-up action carries a named owner + an ISO due date
(smart_actions.validate_register). No anonymous actions, no "ASAP".
Validate against references/QUALITY_CHECKLIST.md — every determination cites
the test it applied; the correct form + the exact deadline; India routes to
hse-india (no national form number); the output is de-identified with <5
suppression and no embedded key; every conclusion traces to evidence.
Assemble the branded report — build report.json (see
assets/regulatory-returns-report.template.json) and run the canonical
report-output call below.
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.
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
Single-threaded by default — a single determination + form-prep is a focused,
tightly-dependent task (the De-identifier scrub still runs FIRST). For a multi-incident
periodic return batch the triage gate fans out to 2 (De-identifier FIRST →
Determination-&-Form-Preparer), then the mandatory review gates. Archetypes:
KB-SNIP-ARCHETYPES.
De-identifier — runs FIRST (sequential gate, not a fan-out peer). A statutory
return holds injured-person identity by law: scrub names, government IDs (SSN /
NI / Aadhaar), contacts, exact dates/locations, and all health detail to stable
role labels in every internal/shared/working copy; apply <5 suppression to any
aggregated summary (e.g. a 300A by department); return the re-identification key
SEPARATELY to the orchestrator — never embed it in the return. Distinguish the
legally-required submission (named, access-controlled) from the circulated copy
(role-labelled) per references/deid-checklist.md. Everything below consumes only
its scrubbed output.
Determination-&-Form-Preparer (India → hse-india) — for the resolved
jurisdiction, apply the cited recordability/reportability test
(KB-DATA-RECORDABILITY-TESTS) to the scrubbed incident facts, return the verdict +
the correct form + the exact deadline + the prepared de-identified return, each
determination traced to its evidence item. For India this job DEFERS to hse-india
via the three-tier degrade: (1) run the relevant hse-india skill
(factories-act-returns / india-accident-notice / india-state-form-finder) as a
subagent; (2) on no-subagent hosts, the main thread reads KB-REG-IN-FACTORIES /
KB-REG-IN-STATEFORMS / KB-REG-IN-PORTALS inline; (3) if hse-india is not
installed, emit routing prose naming factories-act-returns + a KB pointer + a
[GAP]. State detection is mandatory first; it NEVER mints a national form number;
it NEVER hard-blocks. SCOPE-OUT: de-identification (the De-identifier owns it).
Critic/QA (MANDATORY) — adversarial final pass: every determination cites the
test it applied; the correct form + the exact deadline; India routed to hse-india
with no minted national form number; specificity (no reportable yes/no on missing
facts); ZERO de-identification leak (no residual identifier, no <5 cell, no
embedded re-id key). PASS/FAIL.
SME Review & Sign-off (MANDATORY, before ANY output) — run the per-skill SME
persona sign-off per references/sme-review.md (the Regulatory Reporting
Specialist); decision-support that precedes — never replaces — the human
competent-person review.
Simple single-incident determinations run single-threaded — no subagents — but the
De-identifier scrub, the cited-test discipline, and the Critic/QA + SME passes still run.
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:
Build report.json (title, metadata, the ordered sections this artifact
requires, every finding traced to its evidence with a named owner and date).
Resolve branding: the user's brand.yaml overrides the Eyekyam default.
Render both DOCX and PDF from the one report.json via the shared engine.
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.
