The Second Report.

Nobody asked for an assessment. The Production Manager's covering note on the second report said what the first one had said: "Cartons are 12 kg. That's not a heavy lift. It's technique. Re-run the manual handling toolbox talk and get the new starters signed off." The first operator had been on the station for four years. The second was a relief operator in their third week.

A WHS Coordinator has a duty here that is easy to state and easy to skip: the regulations require the risks of hazardous manual tasks to be managed, and a task that has hurt two different people in six weeks has stopped being a training question. Two reports from one station is a finding about the station. The Coordinator did not know yet what was wrong with it. That was the job.

What could be done immediately was to go and look, with the injury reports in hand and a phone in a pocket. What could not be done was to write "technique" a third time.

Line 3 packs bottled product into cartons of twelve. The cartons come off a case sealer onto a short roller conveyor whose discharge end sits at 750 mm. The palletiser stands at the end of it. A floor pallet sits at a right angle to the conveyor, on the operator's right. Each carton is gripped at its ends, no handles, lifted off the rollers, turned through roughly ninety degrees at the trunk and placed in the layer pattern. Seven layers. The bottom layer is placed at 150 mm off the floor, the top at about 1400 mm. When the pallet is full the operator wraps it, pulls it clear with a pallet jack and drops an empty one in.

Both injured operators described the same moment: the bottom two layers, reaching down and across. Neither report recorded that, because the form asks what the person was doing, and both wrote "stacking cartons". The Supervisor, asked how fast the line runs, said "about two a minute, maybe a bit more on the promo pack".

Written as a question the methods can answer:

Is the palletising task within the capacity of the people who do it, and if not, which part of it, the weight, the height, the twist or the rate, carries the risk?

The pallet jack pull is a separate task with a separate method, and this example leaves it at the boundary with a note to come back to it.

The boundary: the palletising station from the conveyor discharge to the finished pallet is in. The pallet jack pull and the empty-pallet fetch are out, noted as a separate push-pull assessment. The roles, as positions: the Palletiser as the primary user; the Line Supervisor and relief operators as secondary; Maintenance, who own the conveyor height and any lifting equipment, as the people the requirements will land on. The modes: normal running; pallet exchange; the promotional pack run, which is a different carton at a higher rate; and the relief operator, whose third week on the station produced the second report.

In ErgoSphere the project is created with the focus set to Manufacturing, which surfaces the physical tools first and hides nothing. The first thing entered is not an analysis. It is the Issues Register: the two injury reports go in with New Issue, one each, the Area set to Line 3 palletising, Safety critical ticked, the report's own words in the Source narrative, the operator's role and not their name, and a status of Open. The register opens before the analysis does, so that whatever the analysis finds is traced back to the two events that started it, and so that the Production Manager's "technique" is on the record as the current control.

The Early Human Factors Analysis wizard was not run, and that is a decision worth recording. It is a concept-stage instrument for a programme, with a standards lens chosen from IOGP 454, AS 7470 or MIL-STD-1472H, and none of those is a packing plant. Its manual handling probe would have said "RULA / REBA", which the Coordinator already knew. A wizard is not run because it is there.

A WHS person usually knows the top of this list better than a Human Factors specialist does, and the bottom of it less well. It is assembled in order of authority, and when two documents disagree the higher one wins.

1. The law and the regulator. The WHS Act and Regulations place a duty on the business to manage the risks of hazardous manual tasks, and they say what makes a task hazardous: repetitive or sustained force, high or sudden force, repetitive movement, sustained or awkward posture, vibration. This task has three of the five before anyone measures anything. The regulations say the risk must be managed. They do not say how to assess it.
2. The Code of Practice. The Code of Practice on hazardous manual tasks is the regulator's own account of how the risk is identified, assessed and controlled, and in most jurisdictions it is admissible as evidence of what is known about the hazard. It is written for exactly this reader. It gives the hierarchy of control. It does not give a number for how heavy, how low or how often.
3. The governing standard. ISO 11228, ergonomics of manual handling, in three parts: lifting and carrying, pushing and pulling, and handling of low loads at high frequency. Part 1 is the one that supplies what the Code does not: reference masses for a population, and the factors that reduce them for height, distance, twist, frequency and grip. Everything on this page that produces a number answers to it.
4. The site's own system. The plant's WHS management system, its hazardous manual task procedure and its incident procedure. These sit under the standard and give it local teeth: they say who signs the risk assessment and who owns a control.
5. Method guides. The HSE's manual handling assessment charts, published as INDG383, are a screening method, not a standard. They enter the register so that the assessment can say which method it used and which edition, and so the reviewer can see that the method was chosen and not just found.

ISO 11228 is in the Standards Register's library, one entry carrying all three parts. Add Standard, From library, and it arrives as a tree: the standard, its three parts as sections, and the clauses under each. Screening makes it the project's, section by section: Applicable, Partial or N/A, with who decided and when, and Confirm to lock it. Part 1, lifting and carrying, is Applicable in full; the carrying clauses stay in because the carton travels a metre in the hands between rollers and pallet. Part 2, pushing and pulling, is N/A with its one-sentence rationale in the box the app marks required: the pallet jack pull is outside this assessment's boundary and will be assessed separately. Part 3, low loads at high frequency, is N/A for a reason the standard itself gives: the carton is well above the mass that part is written for. Both exclusions are dated and attributed, and both appear in the Applicability Statement, so a reviewer can see that the Coordinator decided them rather than missed them.

HSE INDG383, the manual handling assessment charts, is also in the library, and it goes into the register in a different role from the standard. Its three sections are the three operations it covers: Lifting, Carrying and Team handling. Lifting is Applicable. Carrying is N/A, because the carton moves one metre in the hands and the chart's carrying operation is written for a walk. Team handling is N/A, because one person does the task. The register shows the standard type as it is, prescriptive guidance from a regulator in another country, and the professional judgement on the Lifting section says the one thing a reviewer needs: "Method guide adopted for screening under ISO 11228-1; the charts supply a structured walk of the risk factors, not a limit."

The Code of Practice and the site's own procedure are read and held in the Document Register, not entered clause by clause. They are cited in the judgement of the clauses they informed. Nothing that shaped a decision goes unrecorded, but not everything read needs a register entry.

• Is there something here? Before a day is spent on it, does what the Coordinator saw at the station look like a posture problem at all? Cheap, fast, and allowed to be wrong in one direction only.
• Structure. What does the operator actually do, step by step, so that every method reads the same task and the finding can point at a step rather than at "stacking cartons"?
• Posture. At the worst step, how far from neutral is the body, under what load, and does that put the task in a band that requires change?
• Rate. How many lifts per minute, really, for how long a stint? Counted, not asked.
• Capacity. Given the height, the reach, the twist, the rate and the grip, what could a person be expected to lift here, and how far over that is 12 kg?
• Which factor. The question from 1.2. Two methods that name the same driver beat one that has to be argued for.

And one constraint the Coordinator was honest about: this was a WHS person's assessment, done between other duties, on a task that runs eight hours a day. The methods had to be ones a competent generalist can run and defend, and the data had to come from a shift that could be observed in an afternoon.

Two more methods were planned and not applied, and each is recorded with its reason. ART, the HSE's upper-limb repetitive task tool, because the risk this task presents is to the back and not the arms, and an ART score would have been a second number without a second finding. RAPP, the HSE's push-pull tool, because the pallet jack pull is a real task with a real risk and it sits outside this boundary; it is written into the plan as the next assessment, with a date.

ErgoCompass starts with two plain questions: which industry, Manufacturing, and what kind of problem this mainly is, Physical demands on the body, which opens one more, the primary physical demand, Lifting / lowering. The problem statement goes in as the Coordinator wrote it at 1.2. The complexity tenets are rated low: one role, one station, an observable task, a known population. Answering Physical matters for more than the extra question: it moves the tool onto a different pool. The cognitive route ranks against Holman's set of predictive cognitive and systems methods, which contains no physical assessment tool at all; the physical route ranks against ErgoSphere's own physical catalogue, which is where REBA, the NIOSH equation and the MAC live. The palette puts those three at the top with suitability scores of 92, 95 and 88 per cent, and the foundations check asks the question a WHS person most often skips, whether there is a task decomposition under the assessment. There will be, at 4.2. The justification record carries the selected methods, why they fit, and Considered and set aside, which is the four forks above, RULA and ErgoDesign named with their reasons.

Send to HFIP writes the justification into an Integration Plan. On a job this size the plan is two pages, and that is fine; the Planned and not applied section carries ART and RAPP with their reasons, and the RAPP line carries a date, so the pallet jack is a decision on the record and not a gap.

The Coordinator took one photograph from a few metres back and a little behind the operator, as a mid-layer carton went down onto the part-built pallet. Whole body in frame, the carton in the hands, the pallet to the operator's right. That is the whole of the field work at this step, and it took less time than reading this paragraph. Note what the shot is not: it is not the bottom layer at 150 mm, which is the worst lift of the seven and the one the tape measurements and both NIOSH runs are taken on later. The photograph raises the flag. It is not the measurement.

Back at the desk the photograph goes into ErgoLens. The tool has four stages, Capture, Review, Analysis and Report. On the Capture step, Import, then Import Image; a still photograph is accepted alongside video and a live camera. The pose model draws its landmarks on the figure and the pill reads Pose ready. The two risk tiles at the bottom of the Capture step still read "Not scored yet" at this point, which is correct and worth noticing: a still has a pose but no score until the Review step applies a worksheet to it. On the Review step the tabs are Frame Score, RULA and REBA. There is no NIOSH tab on a still, and that is the right behaviour rather than a gap: a lifting equation needs a start and an end, and a photograph has neither, so the tab appears for video captures only.

The frame comes back Very high. RULA is 7 on both sides, the top of its scale; REBA is 5 on both sides, which is its medium band. Two worksheets over one posture, disagreeing about how bad it is, is normal and it is informative: RULA weights the neck and arms, which this frame has in abundance, and REBA weights the trunk and legs, which from this angle it can barely see.

Now the part that matters, and it is the tool refusing rather than the tool scoring. The trunk rotation dial reads n/a, greyed out, with the reason on screen: rotation about the spine's long axis cannot be recovered from a single 2D camera view. The Coordinator had watched the operator turn ninety degrees to the pallet on every carton. The lens could not see it, and it said so instead of reporting zero. Read the dial beside it and the same limitation shows up a second time: trunk flexion 0.0°, from a camera standing behind a man who is plainly leaning into the pallet. Forward lean is out of the image plane from that angle, so the lens cannot measure it either. The instrument under-read this task twice over, in the two channels that carry most of its risk, and it still came back Very high.

The finding is logged from the frame with Log finding, at Severity Medium, with a title that says what it is: "Trunk rotation cannot be measured from this view". Medium, not High, because a lens estimate from one oblique frame is a reason to look, not a result, and because the number it produced is provably the lower bound. The manual REBA at step 4.3, taken on the bottom layer with the twist observed rather than assumed, comes out at 10.

The afternoon starts with a task analysis, and a WHS person's first instinct is to ask why. The answer is that every method that follows needs to point at a step. "Stacking cartons" hurt two people; "2.4, lower and place in the bottom layer, 150 mm, turned ninety degrees" is a step a control can be written against. The tree also does something the injury form could not: it separates the lift, the turn and the placement, which the equation treats as different multipliers, and it puts the pallet exchange and the pallet jack where they belong, at the boundary.

The data came from one observed stint and ten minutes with the operator afterwards, which is where the plan came from. The operator knew the layer pattern and the moment the top layers get easier, and said the promotional carton is the same weight in a bigger box, which is a grip finding nobody had written down. Decompose where the risk lives and stop where it does not: the pallet preparation and the wrap stop at sub-task, the carton placement goes all the way down.

Twelve operations at the leaves, ten of them in boundary, and the risk lives in two of them: 2.3 and 2.4. In ErgoSphere the tree is built on the Tree view with Add Child and Add Sibling, the numbering is automatic, and the Diagram view beside it draws the same structure without editing it, and each node carries its Plan, its Performer, its Preconditions and its HF Notes. Every node also carries a Timing card, Estimate, Min and Max in seconds, a Basis and a Source, and at this point every one of them reads Not timed. That is deliberate. The times arrive at 4.4, observed, and the card will say so.

REBA is done standing beside the station, watching the operator place the first two layers of a fresh pallet, and it is done on the worst moment, not the average one. The posture at 2.4 on the bottom layer: trunk flexed past sixty degrees and turned; neck slightly flexed; both feet on the floor with the knees between thirty and sixty; upper arms reaching forward over the pallet edge at about fifty-five degrees and unsupported; forearms inside the neutral band; wrists near neutral. Then the three modifiers: the load is over 10 kg, the coupling is fair at best on a carton with no handles, and the activity is repeated more than four times a minute, which the next step will confirm.

In ErgoSphere the REBA editor is a wizard in the order the method is scored: Neck Position, Trunk Position, Legs & Support, Upper Arm Position, Lower Arm Position, Wrist Position, then Load, Coupling & Activity, then Results. The SIDE group on the ribbon is set to Right, the side the pallet is on, and the score strip along the top fills as each step is answered. On the load step the bands are the method's own: Over 10 kg (+2), coupling Fair, and Static or repeated activity. The results page gives the Group A score for neck, trunk and legs, the Group B score for arms and wrist, the Final REBA Score, and the risk band with its action. Beside the steps there is a Reference images panel, and the photograph of the placement goes in there, so the posture that was scored travels with the score rather than living in somebody's phone. The same panel exists on the NIOSH editor and takes the tape measurements at 4.5. It is worth knowing that these images are stored beside the project rather than inside it, so they stay on the assessor's machine and do not travel in a shared project file.

The photograph's REBA frame score had been 5, on a view that could measure neither the twist nor the forward lean. The manual assessment, on the bottom layer with both observed rather than estimated, came out at 10. That gap of five points is the whole argument for step 4.3 existing: the flag was right, and it was also the floor. The trunk was the reason in both passes. Finalise and sign marks the assessment complete and writes a compliance snapshot the Requirements workspace can read at stage 7; a finalised REBA is a record, and an unfinalised one is a draft.

The equation at 4.5 and the charts at 4.6 both want a frequency, and the Supervisor's "about two a minute" is the number the Coordinator most needed not to trust. A time and motion observation answers it, and it answers something else on the way: how much of the operator's time is spent waiting, standing at the discharge end with nothing in the hands, which is the number the Production Manager will ask about when the rate comes up.

ErgoTempo runs an observation session. The HTA's placement operations become task cards, Add Task for each, up to twenty: Wait, Grip, Lift and turn, Lower and place, Square, and one for the layer sheet and wrap together. Setup mode to lay them out, then Record mode to start the session, and from there it is one tap per switch: the operator's hands go to the carton, tap Grip; the carton leaves the rollers, tap Lift and turn; it lands, tap Lower and place. Pause All at the pallet exchange, End Session at twenty minutes. End Session goes straight to the Summary, which is a page of its own rather than a tab: Session Information with the linked HTA node, Summary Statistics reading Total Session Time 00:20:00, Active Time 00:19:59, Total Switches 355 and Avg Task Duration 00:00:06, a Session Timeline that draws simultaneous tasks on separate tracks, and the Task Breakdown table with each task's Total Duration, Percentage, Switches and Avg Duration, sorted longest first.

Eighty-three cartons in twenty minutes is 4.2 lifts per minute, twice the Supervisor's figure, and the stint runs the full shift with a break at the pallet exchange. Those are the two inputs the equation was waiting for: frequency and duration. The forty-one percent waiting is not idle time in any sense that matters to the back; it is standing, and it is the reason a rotation to a seated task counts as recovery.

ErgoTempo does not write into the HTA, so the last thing done on this step is done by hand, and the page says so plainly. The averages from the Task Breakdown are typed onto the HTA nodes' Timing cards: 2.3 Lift and turn at 2 seconds, 2.4 Lower and place at 3 seconds with a range of 2 to 5, each with Basis "ErgoTempo observation session, 20 min, 83 cycles" and Source set to Observed. From now on anyone who opens the tree sees a measured time and the session it came from, not a guess with the same typeface as a fact.

The equation starts from a reference mass, the load constant, that a healthy working population could lift under ideal conditions, and reduces it by a multiplier for each way this lift is not ideal: how far the hands are from the body, how low or high they start, how far the load travels, how far the trunk turns, how often the lift repeats and for how long, and how good the grip is. The result is the recommended weight limit for this lift, and the Lifting Index is the actual load divided by it. Under 1 the lift is within the limit for most of the population; between 1 and 3 some workers are at risk; over 3, most are.

The measurements are taken at the station with a tape: horizontal distance from the ankles to the hands, 450 mm on the bottom layer because the pallet edge keeps the operator back from the placement; the two ends of the travel at 150 mm and 750 mm; asymmetry 60 degrees, the trunk turn from the conveyor to the pallet; frequency 4 per minute from 4.4, for a long duration; coupling fair. In the editor these go in as centimetres, not millimetres, which is the unit every distance field on the page is labelled with.

One thing about this lift needs saying before the numbers, because a reader who checks the session will notice it. Taking a carton off a 750 mm conveyor and putting it down at 150 mm is a lowering, and the equation is written for lifts. A lowering is assessed as the lift that reverses it, so the low end goes in as the origin: origin 150 mm, destination 750 mm. That is not a clerical detail. The vertical multiplier is taken from the origin, so entering the conveyor as the origin instead returns a vertical multiplier of 1.00 and reports that the height of this task costs nothing, on the one lift where the height is the reason the lift table gets bought.

In ErgoSphere the NIOSH editor walks those in order: Load Information, with the load weight and the duration of the lifting task set to Long Duration; Horizontal Distance; Vertical Position, with Vertical Origin and Vertical Destination as separate fields so the travel is derived rather than typed; Asymmetry Angle; Frequency, in lifts per minute; Coupling Quality; then Results. The results page shows the Lifting Index on a gauge, the recommended weight limit against the load, and a tile the page is named for: Capacity erosion, LC to RWL, each multiplier drawn as a bar so the reader sees which factor took the capacity away.

Read the multipliers before the index. The load constant of 23 kg has been cut to 3.1 kg on the bottom layer, and the frequency multiplier was the single largest cut. The reach was the next. The twist and the height each took a further fifth. The weight of the carton does not appear in the multipliers at all; it appears only at the end, as the numerator of the index, and 12 kg over 3.1 kg is 3.88. The Production Manager was right that 12 kg is not a heavy lift. The equation says this station makes it one. The top layer, which nobody complained about, is at 3.4: a shorter reach, a worse height, the same rate. Neither layer is acceptable, and the one the site thought was fine is barely better than the one that hurt two people.

The MAC charts walk a lifting task through a fixed list of factors and colour each one green, amber, red or purple, then add the colours' scores to a task score. They were written by a regulator for its inspectors and for exactly this reader, and they ask about things the equation does not: the floor, the environment, whether the posture is constrained by the workplace. They are run here as the second method, on the same lift, so that two framings can be compared.

In ErgoSphere the MAC editor asks for the operation type first, Lifting / lowering, eight factors, and the pre-assessment checklist that decides whether the charts are enough or a full risk assessment is required on top. Three of its fifteen items are true here, each traceable to something already on the record: the carton travels from 150 mm to about 1400 mm, the rate is imposed by the process, and the task has a history of manual handling incidents. The results page then carries a banner saying a full Manual Handling Operations Regulations risk assessment is also required, and names how many items were ticked. Then each factor, with the chart's own wording: load weight and frequency read off the graph, hand distance from the lower back, vertical lift zones, torso twisting and sideways bending, postural constraints, grip on the load, floor surface, environmental factors. The results page gives the Task score, the band strip as G / A / R / P, the drivers, and a sentence the Coordinator read twice: INDG383 publishes no Task-score thresholds. The Task score is for cross-task comparison; per-factor drivers, Red and Purple, are the focus for improvement.

A WHS person calls these controls, and the Code of Practice orders them: eliminate, then engineer, then administer, then protect. A requirement is the same thing written so a fitter can build it and a Coordinator can check it: who, shall, what, under which condition, to what criterion, verified how, with the analysis row as its source. Each of the reds and the big multipliers becomes one. The rate, which the assessment cannot fix, becomes an issue with an owner rather than a requirement nobody can meet. And the retest becomes a requirement of its own, because a control that is installed and never re-measured is a hope.

Every requirement can be failed. Every one traces back to a multiplier, a chart factor or a REBA element, and forward to a way of checking it and a person who checks. None says "safe" or "appropriate". SYS-04 is written knowing the carton is the customer's and will probably come back Not Met, with the reason recorded; it is not deleted for that. And the set is one set, in one register, in the hierarchy's order: the three engineering controls first, the administrative ones after, and no PPE at all, because there is none that lifts a carton.

The HF User Requirements Register holds them. Each is entered with New Requirement and fixed at creation into a requirement group: System, Procedure, Training, Transition or Monitoring, which is the prefix each identifier carries. Then the end-user groups it protects, the Palletiser as primary and relief operators as secondary; the verification method and stage, written as a sentence; an owner, who is Maintenance for the system requirements and the Supervisor for the rest; and a status that starts at Open.

The register carries no clause pointer. It carries a source narrative, a sentence like "NIOSH bottom layer, vertical and distance multipliers; MAC factor C Red", written for the inspector who wants to trace the requirement to the measurement without opening the session. The standard's side of the trace is closed at stage 7 in the Requirements workspace.

The two injury reports were already in the register from 1.3. The analysis added the rest with New Issue: the bottom-layer placement height, the trunk twist at the discharge, the reach across the pallet edge, the carton coupling, the line rate, and the pallet jack pull that was left at the boundary. The physical tools do not raise issues themselves, so each is entered with its source narrative naming the session it came from, the REBA at 4.3, the NIOSH at 4.5, the MAC at 4.6, and linked to the requirement it generated. Each carries a severity, an Area, an owner who can actually close it, a target date, the current control, which is the toolbox talk, and the proposed control.

Status is more than open and closed, and on this job it needed to be. The rate issue went to Transferred once the Production Manager took it onto the production agenda. The carton coupling issue went to Transferred as well, to the customer's packaging specification, with the note that a Not Met is expected. The register never folds Transferred into a closed count, because a risk somebody else now owns has not gone away.

The Part 1 clauses screened in at stage 2 have been waiting in the Requirements workspace. Link Assessment attaches the finalised REBA, the NIOSH sessions and the MAC session to the clauses they answer, the HTA to the task description clause, and the compliance status moves from Not Assessed. At the first pass, before the controls, the lifting clauses are Non-Compliant, with the height and twist issues raised against them and a residual risk rating that says so, and the professional judgement statement says why in one sentence: the recommended limit for the lift as designed is about a quarter of the load. That is an honest first pass, and the workspace is built to hold it. Nothing is marked Not Applicable here for a reason of scope; exclusions belong to screening, and Parts 2 and 3 were excluded there. The four carrying and team-handling clauses inside Part 1 are answered Not Applicable, because Part 1 came in as a whole and those clauses sit inside it, and the task has no carry and no second person.

One number on the header cards is worth explaining before a reader trips over it. The workspace counts 23 designated clauses and reports eight Not Assessed. The fifteen assessed ones are ISO 11228 Part 1, and they are all answered. The eight are HSE INDG383's, and they are deliberately left alone: INDG383 is on this project's Standards Register as the method the MAC assessment comes from, not as a standard the station is judged against. Answering its factors as compliance clauses would re-enter the MAC session's own results in a second place and invite them to drift apart. The governing standard is ISO 11228-1 and the register says so. The assurance score in the corner is low for the same reason, and it is supposed to be at this point in the job: it is a count of answered clauses against a scope that has only just been set.

Five weeks later the lift table was in and the conveyor discharge had been turned to face it. The Coordinator went back with the tape and the phone, and MON-01 was met the way it was written: a REBA and a NIOSH re-run on the same lift at the same rate. The placement now happens at the rollers' height on every layer, with the operator standing close, and with no turn. Horizontal 350 mm, vertical origin and destination both at 750 mm, asymmetry zero, frequency still four a minute for a long duration, coupling still fair.

The engineering controls took the index from 3.9 to 1.7 and did not touch the frequency multiplier, because they could not. The two-hour rotation in PRO-01 moved the duration category and took it to 1.0, on the line between some workers at risk and nominal demand. That is where it stays until the rate changes, and the register says whose decision that is. Both retests are finalised, and the Requirements workspace is revisited: the lifting clauses move to Partially Compliant, with the residual index stated and the rate issue still open against them.

The close-out wizard then builds the assurance report. Each requirement is closed with its evidence: SYS-01 to SYS-03 Verified by the retest, SYS-04 Not Met with the customer's reply on file, PRO-01 and TRN-01 Verified by the roster and the training record, TRA-01 closed as superseded, MON-01 Verified by the retest itself. Every issue is traced to its requirement, and the two injury reports from 1.3 are traced to all of them. Then the four sub-claims, each with a position and a reason: that the programme delivered was the one the risk required, Supported, because the scope came from two injury reports and three instruments were chosen against three named questions; that the requirements were identified and allocated, Supported, because eight were written after the results were read and every one of them can be failed; that issues were managed and closed or transferred, Partially Supported, because the rate is transferred rather than closed and the hand-holds wait on the next packaging revision; that the findings demonstrably influenced the design, Supported, because the lift table and the conveyor turn are the findings, built, and the re-run measures them. The verdict the tool offers for this, and the one selected, is Implemented with conditions, and the conditions are the two Transferred items and the residual index they leave. The completeness check lists them by kind, so issuing with them is a recorded act and not an omission.

The REBA, NIOSH and MAC reports and the assurance summary are marked ready in Deliverables, and Compose Package puts them into one file, hash-sealed, with a manifest, issued as Revision A to the Production Manager and filed against the two incident numbers. The Production Manager opens it in the free Viewer and finds the task tree, the observed times, the three assessments and the argument, and can turn the multipliers tile around without asking for a spreadsheet. The one comment that came back was anchored to the NIOSH retest, and it was the right one: the residual 1.0 depends on a rotation the roster does not yet enforce. Revision B added the roster review to MON-01's evidence, and the comment was resolved against the revision that answered it. The two injury reports were closed in the register on the same day, with their resolution notes pointing at the package.