``` DECISION: SaaS company billing system rebuild (new framework, 9-month timeline) ├─ FRAMING │ ├─ Specific decision under review: Whether to continue committing the remaining $340K/uncommitted budget and runway toward the billing system rebuild, versus pausing/stopping │ ├─ Stage: IN-MOTION RETROFIT — six weeks have elapsed, two of four initial milestones have already slipped, and $160K is already committed; any trigger being defined now is being defined after real evidence of trouble has appeared, not before │ ├─ Trigger authority (proposed): UNKNOWN/DEFAULT — no trigger authority was named at greenlight. By default, the VP of Engineering who sponsored the rebuild and made the "pay for itself within a year" board commitment is the only party with visibility, making them the de facto sole authority — DOCUMENTED as the only party with reporting-line visibility │ └─ Continuation cost in resource terms: $340K of $500K approved budget remains uncommitted (DOCUMENTED); additional cost in elapsed calendar time and team-lead hours already spent "pushing through" two slipped milestones (DOCUMENTED) │ ├─ TRIPWIRE MECHANICS (Phase 3 — all PROVISIONAL) │ ├─ Signal Clarity: UNKNOWN — DOCUMENTED (no criteria defined) — No abandonment criteria were defined at the greenlight meeting; the only observable signals in evidence are milestone slippage (2 of 4), which is real but was never formally tied to any stop/continue decision │ ├─ Lead Time: UNKNOWN — UNKNOWN — No trigger exists, so no lead time can be evaluated; cannot assume any runway between "problem visible" and "point of no return" without inventing a figure │ ├─ Bias Exposure: 4 — DOCUMENTED/INFERRED — The VP has made a public, specific claim to the board ("pay for itself within a year"), creating reputational stake; the VP is also informally pressuring the team lead to "keep pushing through" after slippage, which is direct evidence of sunk-cost behavior already active, not merely a future risk │ ├─ Independent Verification Feasibility: 5 — DOCUMENTED — No one outside the VP's own reporting line has visibility into the rebuild's technical health; this is the most damaging single fact in the evidence — verification is structurally impossible as currently constituted │ └─ Retrofit Contamination: 5 — DOCUMENTED/INFERRED — Applies (IN-MOTION RETROFIT). The only person who could author a trigger today is the VP or people reporting to the VP — the same reporting line that controls all visibility and has already asked the team lead informally to keep going past two slipped milestones. Any trigger written now, by this reporting line, starts maximally contaminated │ ├─ TRIPWIRE CLASSIFICATION │ ├─ Classification: NONE-SET │ ├─ Driving components: Signal Clarity (UNKNOWN — no criteria ever defined) and Independent Verification Feasibility (5 — no outside visibility exists) are the primary drivers; Retrofit Contamination (5) confirms that even a trigger written today would be compromised by the reporting-line structure │ ├─ Rationale: No abandonment criteria were defined at the greenlight meeting, and none have been retrofitted since — the informal "keep pushing through" instruction is not a trigger, it is the opposite: an ad hoc override of the two milestone slips that might otherwise have prompted a stop conversation. There is no dated, numeric, or externally observable condition anywhere in evidence. Because verification is also confined entirely to the VP's own reporting line, even if a trigger were later defined, it could not currently be confirmed by anyone independent — meaning NONE-SET is not just about the absence of a written condition but about the absence of any structural path to detect failure at all. │ └─ Mismatch pattern: NONE IDENTIFIED — no trigger exists to be mismatched; the RETROFIT-CONTAMINATION and NARRATIVE-ESCAPE patterns require an existing trigger to evaluate against, and SILENT-EROSION requires a prior HARDENED baseline, none of which is present in evidence │ ├─ REQUIRED HARDENING ACTIONS (Phase 5) │ ├─ Action requirements: Define an objective, dated or numeric trigger now, before further budget is committed where possible (e.g., a specific milestone/technical benchmark by a specific date, tied to a defined fraction of remaining $340K). Name an independent verifier who is not in the VP's reporting line — someone with the technical access to confirm milestone and health status directly, not via VP-line reporting. State explicitly the lead time this trigger provides before the point of no return (e.g., before the next major budget tranche is committed). │ ├─ Currently satisfied: None of the above — no criteria, no independent verifier, no lead-time statement exist in evidence │ ├─ Currently missing: Objective trigger condition; named independent verifier outside VP's reporting line; lead-time definition; formal (not informal) record of what happens at each future milestone slip │ └─ Independent verifier named: UNKNOWN — evidence states explicitly that no one outside the VP's reporting line has visibility, so no candidate independent verifier can be named from supplied evidence; this itself is a governance gap requiring escalation (e.g., to a CFO, board audit committee, or other function outside Engineering) rather than assumption │ ├─ STABILITY CHECK (Phase 6) │ └─ Result: STABLE — Both forward pass (Signal Clarity → Lead Time → Bias Exposure → Independent Verification → Retrofit Contamination) and reverse pass converge on NONE-SET. The absence of any defined criteria (Signal Clarity: UNKNOWN) and the structural absence of outside visibility (Independent Verification: 5) are each independently sufficient to prevent any classification above NONE-SET, so pass order does not change the outcome. │ ├─ ADVERSARIAL REVIEW (Phase 7) │ └─ Defect found and corrected: Initial read risked treating the VP's informal "keep pushing through" instruction as evidence that an implicit, informally-understood trigger exists (i.e., "the VP is monitoring milestones informally, so something is tracking this"). Corrected: an informal instruction to continue despite slippage is evidence of active sunk-cost pressure overriding any natural checkpoint, not evidence of a functioning trigger — it was reclassified from "weak trigger" to "absence of trigger plus active bias in motion," which is why Bias Exposure was scored 4 rather than left unscored. │ ├─ EVIDENCE GAPS │ └─ UNKNOWN: whether the board's "pay for itself within a year" framing was accompanied by any board-level check-in cadence (this matters because a board check-in, even informal, could function as a partial independent verification point — not present in evidence, so not assumed). UNKNOWN: technical detail behind why the original estimate proved optimistic — relevant to whether milestone slippage reflects a pattern likely to continue or a one-time correction, which would affect Lead Time if a trigger were later defined. │ └─ METHOD NOTE └─ This protocol audits whether a stopping mechanism exists and would survive sunk-cost pressure — it does not judge whether the underlying decision is worth continuing. A HARDENED trigger on a bad decision is still a bad decision, correctly instrumented. VERDICT: PASS ```