| name | weather-dynamic-risk-assessment |
| description | Produce a consultant-grade point-of-work DYNAMIC weather risk assessment for a named renewables site/activity, led by NAMED NUMERIC thresholds -> a hold/stop/evacuate ACTION -> a MANDATORY re-assessment TRIGGER. Use this skill to set weather working limits for turbine/blade/tower or outdoor renewables work, decide when to hold, stop or evacuate for wind, lightning, ice, visibility or temperature, or replace a vague 'monitor the weather' arrangement with defensible numeric triggers. Wind is measured at HUB HEIGHT, not base-of-tower (a base-of-tower reading is a control failure). It OWNS the weather thresholds wind-turbine-work-at-height-rescue (REN-01) only names: BS 7121-1 (2016) 7 m/s man-riding ceiling is the verified anchor, the hub-height cut-off ~=15 m/s is [ASSUMED A4] proposed-and-user-confirms, and every other threshold is user-confirmed or [GAP] -- never invented. De-identifies prior weather-incident context to role labels. Decision-support only; competent-person review required. |
| license | Apache-2.0 |
| metadata | {"author":"eyekyam","version":"1.0","category":"risk-assessment","tier":2,"audience":["M","C"],"industry":["Gen"],"jurisdiction":["UK"],"status":"stable","plugin":"hse-renewables","hse_reviewed_by":"","hse_reviewed_date":""} |
Weather Dynamic Risk Assessment
The hse-renewables point-of-work DYNAMIC weather risk-assessment skill — given a named site / activity and the weather-sensitive task, it produces a defensible weather RA that is led by NAMED NUMERIC thresholds → a hold / stop / evacuate ACTION → a MANDATORY re-assessment TRIGGER (KB-SNIP-DYNAMIC-RA, KB-DATA-WEATHER-THRESHOLDS). Wind is measured at HUB HEIGHT, not base of tower for turbine work — a base-of-tower reading understates the exposure and is treated as a control failure. It grounds the weather-as-a-work-at-height-governor relationship in KB-REG-WAH2005 (weather governs whether work at height proceeds).
The single failure mode this skill exists to kill is "monitor the wind and stop if it gets too windy". A weather artifact with no numeric threshold, no defined action, or no re-assessment trigger is the indefensible paperwork this skill rejects. Every weather parameter (wind, lightning, ice, visibility, temperature) carries a specific numeric trigger → a pre-decided action → a re-assessment event — defined in advance, not decided on the day.
Numeric thresholds are propose-and-user-confirms, never invented. The BS 7121-1 (2016 edition) man-riding ceiling of 16 mph / 7 m/s for a personnel-carrier lift is the one verified public anchor. The hub-height tower-top wind cut-off of ≈ 15 m/s is an [ASSUMED A4] industry baseline — not a single citable standard; it is proposed and the user confirms it (stricter for less-experienced teams). Lightning stand-down follows NFPA 780 practice / a lightning-warning service; manufacturer / CPA in-service crane wind limits, ice-accretion, visibility and temperature triggers are user-confirmed or [GAP]. No threshold is asserted as a fixed standard.
This skill OWNS the weather thresholds REN-01 names (CONV-12, kept distinct). wind-turbine-work-at-height-rescue (REN-01) names the hub-height wind hold / lightning stand-down and defers their ownership here; this skill owns KB-DATA-WEATHER-THRESHOLDS and the dynamic-RA method — cross-referenced, never merged.
When to use this skill
Use this skill when the user needs a point-of-work / dynamic weather risk assessment or weather working limits for a named renewables site and a weather-sensitive activity. Trigger phrases: "set the wind / lightning / ice working limits for blade rope-access on WTG-09", "when do we hold, stop or evacuate for weather", "build a dynamic weather RA for the crane lift / tower-top work", "our method statement says 'monitor the wind' — make it defensible", "what wind speed stops man-riding lifts", "confirm the hub-height wind cut-off". If the request is vague ("keep an eye on the weather" / "stop if it's too windy"), the Workflow intake forces the named site, the weather-sensitive activity, the controlling equipment, and a numeric threshold for each parameter before any drafting. This skill owns the weather thresholds + the dynamic-RA method; the WAH plan + tested rescue is wind-turbine-work-at-height-rescue (REN-01) and lone working is lone-working-assessment (REN-02) — cross-referenced, never rebuilt (CONV-12).
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.
First read the bundle cross-walk ../../knowledge-base/prompt-snippets/renewables-clause-map.md (KB-SNIP-RENEWABLES-CLAUSE-MAP) — it routes the renewables standard → artifact → owning skill (this skill owns the weather thresholds + the dynamic-RA method; REN-01 wind-turbine-work-at-height-rescue only names the thresholds and owns the WAH plan + tested rescue; REN-02 lone-working-assessment owns the lone-working RA). This skill OWNS the weather data + method: read ../../knowledge-base/data-points/weather-thresholds.md (KB-DATA-WEATHER-THRESHOLDS) for the anchors (BS 7121-1 2016 7 m/s man-riding [VERIFIED]; ≈15 m/s hub-height cut-off [ASSUMED A4], proposed-and-user-confirms; NFPA 780 lightning practice; manufacturer/site values [GAP]) and ../../knowledge-base/prompt-snippets/dynamic-ra.md (KB-SNIP-DYNAMIC-RA) for the threshold → action → re-assessment-trigger control spine and the hub-height-not-base measurement discipline. Apply ../../knowledge-base/prompt-snippets/hierarchy-of-controls.md (KB-SNIP-HOC) to every control (eliminate the exposure to weather first — reschedule to a forecast low-weather window / do the work in shelter — before administrative stop-limits). Cross-reference the renewables hazard library ../../knowledge-base/hazard-library/renewables.md for the weather / WAH / dropped-object hazard categories (single home; cross-referenced, never restated). Then resolve the jurisdiction:
| Jurisdiction / element | Read |
|---|
| UK (weather governs whether work at height proceeds — the WAH planning + avoid→prevent→mitigate hierarchy + the duty to not work at height in conditions that endanger) | ../../knowledge-base/regulatory/wah2005.md (KB-REG-WAH2005 — Work at Height Regs 2005 SI 2005/735: reg 4 plan incl. weather, reg 6 avoid→prevent→mitigate, reg 7 collective-before-personal) |
| India (work at height / weather) | ../../knowledge-base/regulatory/in-renewables.md (KB-REG-IN-RENEWABLES — deferral pointer; state detection is mandatory (CONV-8); defer statutory content to the hse-india engine; no national form number invented) |
| Unknown | Ask the jurisdiction / regulatory regime before citing any specific regulation |
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.)
The full typed/branched intake Q-table — the named site / activity (the
specificity anchor: a generic "the wind farm" is refused), the weather-sensitive
task and the controlling equipment (tower-top work / blade rope-access / crane or
man-riding lift / outdoor switching), the weather parameters in play (wind,
lightning, ice, visibility, temperature), the numeric threshold for each parameter
(the core-value gate: a "monitor / stop if too windy" answer with no number is
refused), the measurement point (the hub-height-not-base gate for turbine work),
the action at each threshold (hold / stop / evacuate) and the re-assessment
trigger, and the jurisdiction branch (UK → KB-REG-WAH2005; India →
KB-REG-IN-RENEWABLES + mandatory state detection) — lives in
references/intake.md (the intake-coverage manifest + echo-back +
refuse-on-vague anchors). Run it one question at a time, branch on the answers, and
echo the confirmed site + activity + each parameter's threshold/action/trigger back
before any drafting.
The de-identification step runs FIRST (above): any prior weather-incident
context (a named worker involved in a previous weather-related event) is scrubbed to
role labels before any analysis (references/deid-checklist.md). This is a lower
de-id tier — asset / installation / site-level data dominates — but a clean
role-labelled output is still mandatory.
Then walk the dynamic-RA method (references/METHODOLOGY.md): site / activity /
equipment scoping → for each weather parameter, build the control as a
THRESHOLD → ACTION → RE-ASSESSMENT TRIGGER (KB-SNIP-DYNAMIC-RA): a specific
numeric threshold (the BS 7121-1 2016 7 m/s man-riding figure cited verbatim where a
personnel-carrier lift is used; the ≈15 m/s hub-height tower-top cut-off proposed and
user-confirmed [ASSUMED A4]; lightning per NFPA 780 practice; ice / visibility /
temperature / manufacturer-CPA crane limits user-confirmed or [GAP]) → a pre-decided
action (hold / stop / evacuate) → a mandatory re-assessment trigger (forecast
change, threshold approached, interval). Apply the measurement discipline: wind is
read at hub height, not base of tower — flag a base-of-tower reading used for a
tower-top decision as a control failure. Apply controls.py + KB-SNIP-HOC so the
control leads with eliminating the weather exposure (reschedule to a forecast
low-weather window / shelter the work) before the administrative stop-limit. Re-score
residual risk on the risk_matrix 5×5 after the thresholds + actions, record [GAP]
for any unsupplied numeric basis (manufacturer/site limits, lightning detection radius),
and close each [GAP] with a SMART action (smart_actions) carrying a named role
owner + an ISO due date. Numeric thresholds are user-confirmed / [GAP], never
invented as a fixed standard.
Then: validate the draft against references/QUALITY_CHECKLIST.md → produce the output via the Output format section below. The domain method (the threshold → action → re-assessment-trigger spine + the hub-height measurement discipline + the propose-and-confirm anchor rules) is in references/METHODOLOGY.md.
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
Moderate fan-out for a full dynamic weather RA (the De-identifier runs FIRST as a sequential dependency):
- De-identifier — runs FIRST; scrub any prior weather-incident context (a named
worker / crew involved in a previous weather-related event) and any personal contact
into role labels ("Site Lead (role)", "Rope-Access Supervisor (role)"). This is a
lower de-id tier (asset/site data dominates) but role-labelling is still mandatory.
Returns the re-identification key SEPARATELY to the orchestrator, never to a sibling.
- Threshold-&-Action-Author — for each weather parameter author the
THRESHOLD → ACTION → RE-ASSESSMENT TRIGGER control (
KB-SNIP-DYNAMIC-RA,
KB-DATA-WEATHER-THRESHOLDS): a specific numeric threshold (BS 7121-1 2016 7 m/s
man-riding cited verbatim; ≈15 m/s hub-height cut-off proposed-and-user-confirmed
[ASSUMED A4]; lightning per NFPA 780; ice/visibility/temperature/crane limits
user-confirmed or [GAP]), a pre-decided action (hold/stop/evacuate), and a
re-assessment trigger — measured at hub height, not base of tower; flag a
base-of-tower reading as a control failure. Reject any "monitor the weather / stop
if too windy" control with no number. SCOPE-OUT: does not own the WAH plan / rescue
(REN-01 owns it) or invent a threshold as a fixed standard.
- Residual-&-Actions-Author — apply
controls.py + KB-SNIP-HOC (eliminate the
weather exposure — reschedule / shelter — before the administrative stop-limit), score
residual risk (risk_matrix 5×5) after the thresholds + actions, and write the
owned/dated [GAP]-closure actions (smart_actions) for every unsupplied numeric
basis (manufacturer/site limits, lightning detection radius). SCOPE-OUT: does not set
the per-parameter thresholds (the Threshold-&-Action-Author owns them).
- Critic/QA (MANDATORY) — the renewables-field / dynamic-RA persona
(
references/sme-review.md / KB-SNIP-ARCHETYPES): every weather parameter carries a
numeric threshold → an action → a re-assessment trigger (no "monitor the weather"
control survives); wind is measured at hub height, not base of tower; the BS 7121-1
2016 man-riding figure is cited correctly and the ≈15 m/s anchor is flagged [ASSUMED]
/ proposed-and-confirmed (never asserted as a fixed standard); every other numeric
threshold is user-confirmed or [GAP]; and ZERO prior-incident PII leak into the
circulated copy. Runs the per-skill SME sign-off (decision-support; precedes — never
replaces — the human competent-person review).
Simple single-parameter limits run single-threaded — no subagents.
## 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:
- 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.
