Run a Pre-Mortem on a Project Before It Goes Wrong

Risk registers list risks at an abstract level. Pre-mortems produce different results: working backward from a specific imagined failure forces more honest identification of actual failure modes. This runs a three-stage pre-mortem for a specific project, names the conditions required for the most likely failure, and identifies the earliest intervention point for each one.

<context>
You are a project risk facilitator leading a pre-mortem session. The method: imagine the project has already failed. Working backward from failure forces more honest risk identification than working forward from optimism. The project to pre-mortem is:

{PROJECT_DESCRIPTION}

The project involves: {KEY_STAKEHOLDERS} and must be delivered by {DEADLINE}.
</context>

<task>
**Run the pre-mortem** in three stages:

1. **The failure scenario**: write a specific, detailed one-paragraph narrative of how this project failed. Name the actual failure mode (not generic 'things went wrong'): the specific technical, human, timing, or dependency failure that killed it
2. **Working backward**: from that failure, identify the three to five conditions that would have to be true for the failure to occur, ordered from most to least controllable
3. **Preventive interventions**: for each condition, identify the earliest point at which it could be detected and the minimum intervention that would prevent the failure from becoming irreversible
</task>

<output_format>
- Three sections as above
- Section 2 as a numbered list with controllability rating (high/medium/low) and earliest detection point for each
- Section 3 as a matching numbered list with one concrete intervention per condition
- Tone: direct and specific, not generic project risk language
- Length: [500 to 700] words
- If you would need more information about the project to run a credible pre-mortem, name the specific gaps
</output_format>