Produces a scenario-keyed emergency response plan (ERP) for a named site or facility — call-out tree, muster and evacuation arrangements, scenario response procedures, drill schedule, and external-responder integration. Use this skill whenever a user asks to write or review an emergency plan, evacuation plan, emergency procedures, crisis-response plan, or to plan emergency drills for a specific location. It grounds the plan in the site's credible emergency scenarios and the hierarchy of controls (prevention before response), assigns named roles with deputies, and schedules drills — emitted as a branded report. For continuity of critical activities after the emergency is controlled (RTO/RPO/MTPD) use business-continuity-plan; for mine-rescue-specific arrangements use mine-rescue-erp; for offshore use the marine-emergency-response skill. Decision-support only; a competent person must review the output.
Produces a scenario-keyed emergency response plan (ERP) for a named site or facility — call-out tree, muster and evacuation arrangements, scenario response procedures, drill schedule, and external-responder integration. Use this skill whenever a user asks to write or review an emergency plan, evacuation plan, emergency procedures, crisis-response plan, or to plan emergency drills for a specific location. It grounds the plan in the site's credible emergency scenarios and the hierarchy of controls (prevention before response), assigns named roles with deputies, and schedules drills — emitted as a branded report. For continuity of critical activities after the emergency is controlled (RTO/RPO/MTPD) use business-continuity-plan; for mine-rescue-specific arrangements use mine-rescue-erp; for offshore use the marine-emergency-response skill. Decision-support only; a competent person must review the output.
A consultant-grade HSE skill that produces a scenario-keyed emergency response plan
(ERP) for a named site or facility grounded in ISO 45001 clause 8.2 (emergency
preparedness and response). It forces the single lever that separates a defensible
artifact from copy-paste paperwork: the plan is built from the site's and the
(eliminate/reduce the scenario at source first, then plan the response), never a generic
"evacuate" procedure. Every response role has a ; muster points are
; the (cadence from ); external
responders are with a site-interface sheet.
## Output format
credible
emergency scenarios
hierarchy of controls — prevention before response
named deputy
named
drill schedule is dated
KB-DATA-DRILL-FREQ
integrated
The ERP and business-continuity-plan are adjacent but distinct (D-07): this skill
plans the immediate emergency response (scenarios, muster, evacuation, call-out
tree); for continuity of critical activities after the emergency is controlled
(RTO / RPO / MTPD, recovery strategies) use business-continuity-plan — the ERP
output carries a one-line pointer to it. This generic ERP does not subsume
mine-rescue-erp (the distinct mining-rescue artifact) or the marine /
offshore-emergency skill — route those requests to their own skills.
When to use this skill
Use this skill when the user needs an emergency response plan for a concrete named
site or facility — for example "write an ERP for the Plant 4 solvent-storage
warehouse", "review our evacuation plan for the head-office building", "build emergency
procedures for a chemical-release scenario at the bulk-tank farm", or "plan the fire and
medical drills for the distribution centre". Trigger phrases: emergency response plan,
ERP, emergency plan, evacuation plan, emergency procedures, crisis-response plan,
call-out tree, muster point, emergency drill, emergency preparedness.
Route elsewhere: for continuity of critical activities after the incident is
controlled (RTO/RPO/MTPD, recovery roles) use business-continuity-plan; for a
mine-rescue plan use mine-rescue-erp; for an offshore/marine emergency
plan use the marine-emergency-response skill. The intake's Q1 disambiguates ERP
(scenarios / muster / evacuation / immediate response) from BCP (critical activities /
RTO / RPO / MTPD). If the request is vague ("write me an emergency plan"), the Workflow
intake below refuses to proceed until the named site + at least one credible
scenario + muster/evacuation arrangements are captured.
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.
Jurisdiction / scope
Read
India
../../knowledge-base/regulatory/in-factories-act.md (s.41B on-site emergency plan for MAH installations; + in-state-forms.md for the user's state — defers to hse-india)
UK
../../knowledge-base/regulatory/uk-hswa.md (Regulatory Reform (Fire Safety) Order 2005 art. 15; DSEAR 2002 where flammable atmospheres)
Drill frequencies by scenario / site-class (always)
../../knowledge-base/data-points/drill-freq.md (KB-DATA-DRILL-FREQ — quote source+year; [GAP] where unresolved)
Operations clause cross-walk (always)
../../knowledge-base/prompt-snippets/ops-clause-map.md (KB-SNIP-OPS-CLAUSE-MAP — ISO 45001 8.2 → this skill)
This skill always grounds in KB-STD-ISO45001 (clause 8.2 emergency preparedness &
response) and applies KB-SNIP-HOC so prevention precedes response on every
scenario; it builds the scenario menu and per-scenario procedure skeletons from
KB-SNIP-ERP-SCENARIOS and sets the dated drill cadence from KB-DATA-DRILL-FREQ
(quote source+year; [GAP] where unresolved — never a bare number). For an India
site it resolves the state via the hse-india engine (mandatory state detection — never
a national form number). The bundle clause cross-walk is KB-SNIP-OPS-CLAUSE-MAP. 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 ERP-vs-BCP disambiguation at Q1, the per-scenario branch (Q2), 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, and
refuse to proceed on a vague request (record [ASSUMPTION] / [GAP], never invent).
The six questions:
Q1 — Site & occupancy + ERP/BCP disambiguation(the specificity anchor; refuse
generic). The named site/facility + headcount/shift pattern. This skill plans the
immediate emergency response (scenarios, muster, evacuation, call-out tree). If the
user actually wants continuity of critical activities after the emergency is
controlled (RTO / RPO / MTPD, recovery strategies), route them to
business-continuity-plan — do not produce a generic plan here.
Q2 — Credible scenarios(multi-select, branch per scenario): fire/explosion ·
chemical/gas release · medical · structural/collapse · severe weather/flood ·
security/violence · loss of utilities. Each selected scenario activates a
scenario-specific procedure stub from KB-SNIP-ERP-SCENARIOS.
Q3 — On-site response capability(MCQ): first-aiders only / trained ERT / fire
team / none — capability must be proven before the plan relies on it.
Q5 — Jurisdiction (for the legal ERP baseline; India → mandatory state detection
via hse-india, never a national form number).
Q6 — Drill history(free-text, optional) — seeds the dated drill schedule
against KB-DATA-DRILL-FREQ.
Refuse-on-vague anchor: no plan until the named site + ≥1 credible scenario +
muster/evacuation arrangements are captured. A generic "evacuate" plan with no
scenario-keyed steps is explicitly refused.
Then: apply the ERP method (references/METHODOLOGY.md) — scenario selection →
prevention-first (hierarchy) → response procedure → roles + deputies → drill cadence →
responder integration → validate the draft against references/QUALITY_CHECKLIST.md →
produce the output via the Output format section 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
This is a moderate roster (A6 "moderate = 2–4"): the De-identifier is the
sequential first gate (the call-out tree holds real names + phone numbers — a
legitimate internal operational record, but role + duty-phone labels in any widely
distributed copy). The fan-out is per-scenario for a multi-scenario facility; a
small single-scenario site runs single-threaded. The prevention-first hierarchy is
applied via the A7 controls engine + KB-SNIP-HOC (prevention/elimination before the
plan relies on response), and the dated drill schedule is built with smart_actions
against KB-DATA-DRILL-FREQ — these are deterministic script calls, not fan-out jobs.
Archetypes: KB-SNIP-ARCHETYPES.
De-identifier — runs FIRST (sequential gate, not a fan-out peer); scrub responder
names / personal mobile numbers / health detail in the call-out tree to role +
duty-phone labels for any circulated copy (the internal master may hold the real
contacts as a legitimate operational record). Every job below consumes the scrubbed text.
Scenario-Analyst — for each credible scenario from Q2, build the per-scenario
response procedure from KB-SNIP-ERP-SCENARIOS — prevention controls first (apply
controls + KB-SNIP-HOC), then detection/alarm → immediate actions → escalation. Each
procedure is scenario-keyed, never a generic "evacuate". SCOPE-OUT: roles/muster
(Roles-&-Logistics), drill cadence (Drill-Planner).
Roles-&-Logistics — build the call-out tree & roles with named deputies, the
named muster/assembly points and evacuation routes, and the external-responder
integration sheet (access, isolation points, site interface). SCOPE-OUT: per-scenario
procedures (Scenario-Analyst).
Drill-Planner(optional) — set the dated drill schedule by scenario/site-class
from KB-DATA-DRILL-FREQ (quote source+year; [GAP] where unresolved) as smart_actions
with named owners + dates. Merge into Roles-&-Logistics for a small site.
SME Reviewer (MANDATORY pre-output gate) — runs the Emergency Planning Officer
sign-off in references/sme-review.md before any output: prevention precedes response,
every role has a deputy, muster points named, drills dated, the response capability is
proven (not assumed), and the ERP→BCP boundary is honoured.
Critic/QA (MANDATORY) — adversarial final pass: prevention-before-response (no
response-only plan), scenario-keyed procedures (no generic evacuate), every role
deputised, muster named, drills dated, citations trace to the KB (8.2 / 1910.38), and
ZERO de-id leak (no personal mobile number in a circulated copy). PASS/FAIL.
Simple single-scenario sites run single-threaded — no subagents — but the De-identifier
scrub, the controls/smart_actions calls, and the Critic/QA + SME passes are still made.
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.
