Stalwart Tender Intelligence

Hunt for open tenders

Paste a TenderBoard listing here. Every row is parsed and triaged against your tendering capacity, the closing date, and whether it is guarding work at all. Triage is not an eligibility verdict: a listing line does not carry the tender's stated requirements, so it can only tell you what is worth opening. Use the Eligibility gate once you have the document.

Set your TenderBoard filter first. TenderBoard has an industry category called Facilities Management: Security Services. Filtering to it at source means far less noise reaches this screen. You are otherwise looking at well over a thousand live tenders.

TenderBoard listing text

Pipeline

Everything hunted, soonest closing first. Verdicts are triage only.

Eligibility gate

Paste the tender's stated requirements. This screens them against what COMPANY_STATE actually evidences, and it runs before any costing. Two of the seven documented losses were never pricing losses.

Requirement text from the tender document

What you are being screened against

Read live from the ledger, never hardcoded. If a fact is missing it is treated as not held.

Cost the tender

The same engine as v33, extracted and verified against it to the cent across seven scenarios. Check the roster before you trust a price: a headcount that breaches the PRD hours ceiling is not a saving, it is licence exposure.

Security manpower

Contract and statutory

PWM start year

Term (years)

CPF age band

Target margin %

Annual leave (days)

Sick leave (days)

Reservist (days)

Parental (days)

Hospitalisation (days)

KPI allowance ($/officer)

Include 11 public holidays in the relief calculation

Equipment, annual cost

These start empty deliberately. An unset line exports at $0, which is how mandatory body cams were priced at nothing on Grange Heights. Enter the real figure or tick N/A.

Roster legality

An ordinary month is 44 hours a week, so 190.67 hours. PRD allows a maximum 72 extra hours, giving a hard ceiling of 262.67 hours per officer per month. A 12-hour, 26-day roster is 312 hours and cannot lawfully be delivered. Run this before pricing.

Pattern

Shifts per year

Shift length (hours)

Posts at 24/7

Officers quoted per post

Relief buffer %

Liquidated damages and the termination trigger

The termination clause is reported separately from margin erosion on purpose. When it fires first, no pricing decision defends the contract and roster discipline is the binding risk.

Negotiation floor %

High risk %

Terminate after N incidents/month

Client and internal documents

Every figure comes from the costing engine, so an exported document cannot disagree with what is on screen. Layouts are the same as v33, so these look like the documents already sent. The logo is embedded, which fixes the broken image v33 still produces.

Client name

Site / location

Attention to

Prepared by

Executive summary, optional

Preliminary risk assessment, optional

Manager rationale, appears in the assessment

Download

The tender assessment and the intro letter are never blocked. The assessment is the bid or no-bid pack, and that decision usually has to be taken before equipment is costed, so instead of refusing it records what was missing on its own last page. The quotation and the internal costing are gated: an unset equipment line would export at $0, and that is how mandatory body cams were priced at nothing on Grange Heights.

Pre-submission Claim Check

Paste the bid's compliance and capability wording. Every rule here traces to a real defect found in a real Stalwart submission. Run it before the document goes out, not after.

What this caught on its first run. The AirTrunk certifications table claimed ISO 27001, PCI-DSS and SOC 2. Stalwart holds none of them. The same bid called a S$20M insurance upgrade “confirmed” a month before any quotation existed. Both went out.

Bid text

What Stalwart can actually evidence

Read from the ledger. Anything not on this list is written as provisional, never as held.

The ledger

Every figure the system is allowed to use, with where it came from, when it was last verified and who confirmed it. Anything past its review date is marked unverified rather than quietly presented as current.

Derived facts

Conclusions, stored with the rule and the inputs that produced them. A conclusion whose inputs have moved marks itself stale instead of staying quietly wrong. This table exists because “Stalwart qualifies for a 3-year GSR registration” survived for months as a bare conclusion.

Who has access

Everyone here signs in with their own email and the shared passphrase. That reliably keeps out anyone not on this list, and every change to the ledger is recorded against a name. It does not prove which of you signed in, because the passphrase is shared. If someone leaves, remove them here and change the passphrase, because they still know it.

Add someone

Work email

Name

Role

Access log

Sign-ins, refusals, and every change to this list. Refusals matter: repeated ones from an address you do not recognise are worth knowing about.

Regulatory watch

Five official sources, checked daily at 06:00 Singapore time. The watcher never writes a figure. It raises what moved and links the source; you confirm before anything reaches costing.

Changes awaiting your decision

Confirm means the source really moved and the ledger needs updating. Dismiss means it was noise.