borghei/Claude-Skills / project-management/discovery/pre-mortem
pre-mortem
VerifiedRuns a pre-mortem that sorts launch risks into evidenced tigers, paper tigers and unspoken elephants, each with an owner. Use when a big launch, build or migration is coming, or the team has a bad feeling it cannot name.
Pre-Mortem Risk Analysis Expert
Overview
A pre-mortem is a prospective hindsight exercise: imagine that your product has launched and failed, then work backward to identify why. This skill uses the Tiger / Paper Tiger / Elephant classification to categorize risks by type and urgency, ensuring launch-blocking issues are addressed before launch while avoiding wasted effort on unlikely risks.
Core Capabilities
- Prospective hindsight framing — "It is 14 days after launch, and we failed. Why?" surfaces specific, honest risks.
- Tiger / Paper Tiger / Elephant classification — separate real evidenced risks (Tigers) from anxiety (Paper Tigers) and unspoken concerns (Elephants).
- Urgency triage — tag each Tiger Launch-Blocking, Fast-Follow, or Track with owner and decision date.
- 6-phase facilitation — scene-setting, silent generation, cluster, classify, mitigate, address elephants in 60-90 minutes.
- Automated categorization —
risk_categorizer.pysummarizes distribution and flags elephants needing escalation.
When to Use
- Before committing significant resources to build (post-ideation, post-validation).
- Before a major launch, migration, or architectural change.
- When the team has "a bad feeling" they cannot articulate.
- When stakeholder confidence is high and you need to stress-test it.
Clarify First
Before running the pre-mortem, settle the inputs below. When one is missing or vague and would change the result, ask about it; when it wouldn't, use a sensible default and say which one you used.
- The thing being stress-tested: the specific launch/build/migration that "failed 14 days after launch" (anchors the prospective-hindsight framing).
- Failure horizon: the date/milestone you imagine looking back from (a vague horizon produces vague risks).
- Who's in the room: the 4-8 participants and whether psychological safety exists (elephants only surface with candor and the right people).
Ask about the two or three inputs that matter most for this request, not the whole list. If the user wants a result straight away, produce it and state your assumptions at the top.
Workflow
- Settle the Clarify First inputs and run the session from
references/methodology-and-tools.md(scene-setting, silent generation, cluster, classify, mitigate, elephants). - Record each risk with
description,category(tiger/paper_tiger/elephant) andevidence; tigers also needurgency(launch_blocking/fast_follow/track), and launch-blockers anowneranddecision_date. - Run
risk_categorizer.pyand fix its findings before the session closes. - Done when the tool exits 0, every launch-blocker has an owner and a decision date, and each elephant has a person who will raise it.
python3 project-management/discovery/pre-mortem/scripts/risk_categorizer.py project-management/discovery/pre-mortem/assets/sample_risks.json # exit 0
python3 project-management/discovery/pre-mortem/scripts/risk_categorizer.py project-management/discovery/pre-mortem/assets/sample_risks_flawed.json # exit 1
python3 project-management/discovery/pre-mortem/scripts/risk_categorizer.py --demo --format json
The tool groups risks by category and urgency, flags elephants whose text carries escalation signals (no owner, blocked, security, compliance, legal and similar), and reports two findings a rushed session skips: a launch-blocking tiger with no owner or decision_date, and a tiger whose evidence is speculative ("might", "could", "we think"), which makes it a paper tiger until shown otherwise. Exit codes: 0 no findings, 1 findings, 2 bad input (missing file, invalid JSON, missing fields). Document the session with assets/pre_mortem_template.md.
Anti-Patterns
Mitigating Paper Tigers
Mistake: "AWS regional outage on launch day" is classed as a Tiger and gets weeks of redundancy work, with no evidence that it is likely.
Why it happens: A frightening consequence feels like a high probability.
Instead: For every Tiger, write the evidence that it could actually happen. A described consequence alone does not count. With no evidence, reclassify it as paper_tiger and cover it with a standard runbook.
Risks too vague to act on
Mistake: The list holds "stakeholder issues", "scope creep" and "technical risk". Why it happens: Generic labels are safe to say aloud and need no evidence. Instead: Describe each risk as a concrete failure scenario in 1-2 sentences, such as "the CRO blocks launch on Day -7 because the enterprise SKU was cut". Translate worries ("the design feels weak") into a measurable consequence, or drop them.
Running the pre-mortem too late
Mistake: The session happens a week before launch and surfaces a Tiger that needs four weeks of work. The team launches anyway. Why it happens: Pre-mortems get scheduled as a launch-readiness check instead of a planning step. Instead: Run the first pre-mortem at project kickoff and a second one around Day -30. For each risk, confirm it can still be mitigated in the time left.
Nine more, each with a bad and a good example: references/red-flags.md.
References
Load the reference that matches the task — keep this file lean and pull detail on demand:
- references/methodology-and-tools.md — the thought experiment, full Tiger/Paper Tiger/Elephant classification with examples, urgency table, the 6-phase facilitation script,
risk_categorizer.pyusage and flags, output formats, troubleshooting, success criteria, and bibliography. Read when running or scripting a session. - references/pre-mortem-guide.md — Gary Klein's origin and why pre-mortems work, when (and when not) to run one, deep classification criteria across domains, the facilitation timeline, and a worked SaaS launch example. Read for the theory and a full example.
- references/red-flags.md — 12 anti-patterns (mitigation theater, groupthink, vague risks, over-mitigated elephants, late pre-mortems, missing owners) with bad/good examples and one-line checks. Read before sharing pre-mortem output.
Scope & Limitations
In Scope: prospective hindsight using the "14 days after launch failure" framing; Tiger / Paper Tiger / Elephant classification with urgency levels; automated elephant escalation detection; risk registry generation; facilitation methodology for in-person and remote teams.
Out of Scope: ongoing risk management and tracking (see senior-pm/scripts/risk_matrix_analyzer.py); quantitative probability/impact scoring (senior-pm/); product discovery and hypothesis validation (brainstorm-experiments/); technical architecture risk assessment (engineering/ skills).
Important Caveats: most effective with 4-8 participants; the elephant escalation keyword check is a heuristic; pre-mortems complement, not replace, assumption mapping (identify-assumptions/); psychological safety is a prerequisite.
Files
| File | Purpose |
|---|---|
scripts/risk_categorizer.py | Categorizes risks, flags elephants to escalate, gates on unowned launch-blockers and speculative tigers |
assets/sample_risks.json, assets/sample_risks_flawed.json | Runnable inputs: sound (exit 0) and flawed (exit 1) |
assets/pre_mortem_template.md | Session template |
examples/shared-dashboards-v1-pre-mortem.md | Worked session |
references/*.md | See References |
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution. | Stars: 114k | Updated | Claude CodeCodex | Verified | Install |
| add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation. | Stars: 114k | Updated | Claude CodeCodex | Verified | Install |
| diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope. | Stars: 114k | Updated | Claude CodeCodex | Verified | Install |
| fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks. | Stars: 114k | Updated | Claude CodeCodex | Verified | Install |