Designs and runs a consultant-grade leadership safety walk / gemba walk (genchi-genbutsu "go and see") for a named area, task, or shift. Use this skill whenever a leader, manager, or supervisor wants to run a felt-leadership safety walk, a gemba walk, a leadership site visit, a safety conversation or engagement walk, or a worker-engagement / felt-leadership round — NOT a tick-box safety inspection or audit. It builds the walk from OPEN conversation prompts (engagement, never a closed yes/no checklist), captures worker concerns role-labelled to protect psychological safety, and converts every commitment made on the walk into an owned, dated smart_actions action tracked to closure — the count and closure-rate of which is itself a leading indicator. Grounded in ISO 45001:2018 clauses 5.1 (leadership and commitment) and 5.4 (consultation and participation of workers) and HSE HSG65 felt leadership. Decision-support only; a competent person must review the output.
Designs and runs a consultant-grade leadership safety walk / gemba walk (genchi-genbutsu "go and see") for a named area, task, or shift. Use this skill whenever a leader, manager, or supervisor wants to run a felt-leadership safety walk, a gemba walk, a leadership site visit, a safety conversation or engagement walk, or a worker-engagement / felt-leadership round — NOT a tick-box safety inspection or audit. It builds the walk from OPEN conversation prompts (engagement, never a closed yes/no checklist), captures worker concerns role-labelled to protect psychological safety, and converts every commitment made on the walk into an owned, dated smart_actions action tracked to closure — the count and closure-rate of which is itself a leading indicator. Grounded in ISO 45001:2018 clauses 5.1 (leadership and commitment) and 5.4 (consultation and participation of workers) and HSE HSG65 felt leadership. Decision-support only; a competent person must review the output.
A consultant-grade HSE skill that designs and runs a leadership safety walk / gemba
walk (genchi-genbutsu — "go and see") for a named area, task, or shift, grounded
in ISO 45001:2018 clauses 5.1 (leadership and commitment) and 5.4 (consultation
and participation of workers) and (,
). It forces the single lever that separates an engagement
walk from copy-paste paperwork: the walk is built from that
invite the worker's own account — (a closed
checklist is an audit, not a gemba walk, and is for specificity/defensibility).
Worker concerns are captured to protect psychological safety — . And every the leader
makes on the walk becomes an —
the
(). A walk that produces no tracked commitment fails the
defensibility gate: a walk is not a conversation that evaporates, it is commitments owned,
dated, and closed.
## Output format
HSE HSG65 felt leadership
KB-SNIP-GEMBA-PROMPTS
KB-SNIP-LEADERSHIP-CLAUSE-MAP
OPEN conversation prompts
never a closed yes/no tick-box checklist
flagged
role-labelled
no
concern is ever attributed to a nameable individual
commitment
owned, dated smart_actions action tracked to closure
count and closure-rate of walk commitments is itself a leading indicator
KB-DATA-LEADING-INDICATORS
When to use this skill
Use this skill when a leader, manager, or supervisor wants to run an engagement safety
walk for a named area / task / shift — for example "design a felt-leadership safety
walk for the night-shift warehouse pick-face", "run a gemba walk on the line-2 changeover
and capture what the operators actually deal with", or "I'm doing a leadership site visit
on the scaffold deck — give me open prompts and turn my commitments into tracked actions".
Trigger phrases: safety walk, gemba walk, genchi-genbutsu, felt-leadership walk / round,
leadership site visit, safety conversation / engagement walk, worker-engagement walk,
walk-and-talk. This skill is not a tick-box safety inspection or audit (that is a
safety-audit), not a behaviour-observation programme (that is bbs-program-designer),
not a maturity survey (that is safety-culture-assessment), and not a KPI
framework (that is leading-lagging-kpi-framework) — intake Q1 disambiguates the walk
from these siblings. If the request is vague, or asks for a closed yes/no checklist, the
Workflow intake below refuses to proceed and steers to open, area-specific prompts.
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.
This skill always grounds the walk in the gemba prompt bank, the leadership clause
map, and the leading-indicator catalogue, then resolves the jurisdiction only for any
consultation/participation legal duty:
Read on every run
File
The open-question gemba prompt bank by purpose/area (felt-leadership/engagement · hazard-spotting · control-verification · post-incident) — engagement, never a tick-box; concerns role-labelled; commitment-closure as a leading indicator
The ISO 45001 leadership clause cross-walk — this skill owns clauses 5.1 (leadership and commitment, felt leadership) and 5.4 (consultation and participation of workers)
The leading/lagging indicator catalogue (single-sourced) — read the gemba-commitment closure-rate entry; report walk-commitment closure as a leading indicator, quote source+year
../../knowledge-base/regulatory/in-factories-act.md (safety-committee / worker-participation provisions; defers to hse-india; mandatory state detection, never a national form number)
Unknown
Ask before citing any specific law — the gemba method (open prompts, felt leadership, tracked commitments) still applies
Always apply KB-SNIP-HOC to any control a commitment implies. For the
commitment-closure leading indicator, resolve the ID in KB-DATA-LEADING-INDICATORS
and quote its source+year — never a remembered figure. 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 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, and refuse to proceed on a vague request. The hard anchors:
Q1 (purpose + which skill) — confirm this is an engagement walk, and
disambiguate it from the siblings: a tick-box inspection/audit → safety-audit;
a structured behaviour-observation programme → bbs-program-designer; a maturity
survey → safety-culture-assessment; a KPI framework → leading-lagging-kpi-framework.
If the user asks for a closed yes/no checklist, refuse — a gemba walk is open prompts,
not a tick-box, and steer them to the open prompt bank (or to safety-audit if they truly
want an inspection).
Q-subject (named area / task / shift) — the specificity anchor. Refuse a vague
"do a site walk" — name the specific area, task, or shift the walk covers, because
the prompt families are selected per area.
Q-evidence (walk purpose → prompt family) — felt-leadership/engagement ·
hazard-spotting · control-verification · post-incident reassurance — selects the open
prompts from KB-SNIP-GEMBA-PROMPTS.
The gemba-walk method (felt leadership / genchi-genbutsu)
Full method in references/METHODOLOGY.md. Steps:
De-identify the inputs FIRST — any worker concerns, names, or shift/role detail the
user brings are scrubbed to role labels before any drafting (the deid block above +
references/deid-checklist.md + the De-identifier-runs-first orchestration rule). Worker
concerns are captured role-labelled (psychological safety) — never attributed to a
nameable individual. A named-worker→concern attribution is a de_identificationhard-fail.
Select the open prompts by purpose + area — from KB-SNIP-GEMBA-PROMPTS, pick the
prompt family for the walk's purpose (Q-evidence) and anchor each prompt to the specific
area/task (Q-subject). Prompts are open, area-specific, non-interrogative — they
invite the worker's own account. A closed yes/no tick-box checklist is refused and
flagged (specificity · defensibility).
Run the conversation, capture concerns role-labelled — record what workers raise
against role/group labels only; no individual is named. Flag [GAP] where the walk did
not reach an area in scope.
Convert every commitment to an owned/dated action — each commitment the leader makes
on the walk becomes a SMART action (specific, measurable, assignable (named
role-label owner), relevant, time-bound (ISO due date)) via
smart_actions.validate_register. No anonymous commitments, no "ASAP". Any control a
commitment implies is ranked via KB-SNIP-HOC (controls), higher-order first.
Report commitment closure as a leading indicator — the count and closure-rate of
walk commitments is a leading indicator (resolve KB-DATA-LEADING-INDICATORS, quote
source+year). A walk that produces no tracked commitment fails the defensibility
gate — it is recorded as [GAP], not signed off as complete.
Validate against references/QUALITY_CHECKLIST.md — open prompts (not tick-box);
every concern role-labelled (no individual named); every commitment owned + dated +
tracked to closure; commitment closure surfaced as a leading indicator; ISO 45001 5.1/5.4
cited accurately; de-id applied.
Assemble the branded report — build report.json (see assets/report.json) and run
the canonical report-output call below.
The orchestration block (below) sits after this Workflow so the triage gate can judge the
assembled walk before deciding to fan out. The De-identifier runs FIRST as a sequential
gate in every case.
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 the moderate roster with the De-identifier as the sequential first gate —
critical here because worker concerns raised on a walk are sensitive: attributing a concern
to a nameable individual breaks psychological safety. Commitment-to-action conversion is a
deterministic smart_actions call at Workflow step 4 — never an LLM fan-out job.
Archetypes: KB-SNIP-ARCHETYPES.
De-identifier — runs FIRST (sequential gate, not a fan-out peer). Before any
drafting: scrub every nameable individual to a role/group label; capture worker
concerns role-labelled (psychological safety); ensure no concern is attributed to a
nameable individual. Apply references/deid-checklist.md in full. Return the
re-identification key SEPARATELY to the orchestrator — never to a sibling, never in the
document. Every job below consumes only this scrubbed output. SCOPE-OUT: does not select
prompts (Walk-Designer) or draft actions (Commitment-Tracker).
Walk-Designer — for the walk's purpose + area (intake Q-evidence / Q-subject), select
the open conversation prompts from KB-SNIP-GEMBA-PROMPTS; anchor each to the specific
area/task; keep them open, area-specific, non-interrogative. Refuse a closed tick-box
checklist. Consumes only the scrubbed output; names roles/groups, never individuals.
SCOPE-OUT: converting commitments to actions (Commitment-Tracker).
Commitment-Tracker — turn each commitment made on the walk into a SMART action
with a named role-label owner + an ISO due date + a measure (smart_actions.validate_register),
and surface the commitment count / closure-rate as a leading indicator
(KB-DATA-LEADING-INDICATORS). SCOPE-OUT: selecting prompts (Walk-Designer),
de-identification (De-identifier).
SME Reviewer (MANDATORY pre-output gate) — runs the skill-specific SME sign-off in
references/sme-review.md (HSE Leadership / Engagement Coach) before any output: open
prompts (not a tick-box), worker concerns role-labelled (no individual named), every
commitment owned + dated + tracked to closure, commitment closure surfaced as a leading
indicator.
Critic/QA (MANDATORY) — adversarial final pass: open prompts not a closed checklist,
every commitment owned + dated + tracked to closure (no untracked commitment), commitment
closure reported as a leading indicator, ISO 45001 5.1/5.4 cited accurately, and ZERO
de-identification leak (no concern attributed to a nameable individual, no re-id key in
the output). PASS/FAIL.
A single short walk with one area may run single-threaded — no subagents — but the
De-identifier scrub, the smart_actions commitment conversion, and the Critic/QA pass 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.
