How the Pre-Mortem Skill Classifies Risks: A Three-Tier Taxonomy Explained

The pre-mortem skill classifies risks into three primary categories—Tigers (real threats), Paper Tigers (overblown concerns), and Elephants (unspoken worries)—then applies a secondary urgency tag to Tigers designating them as Launch-Blocking, Fast-Follow, or Track based on mitigation timelines.

The pm-skills repository provides structured product management methodologies as executable skills, with the pre-mortem analysis being a critical component for launch risk assessment. When you invoke the /pre-mortem command, the skill applies a specific two-level classification system defined in pm-execution/skills/pre-mortem/SKILL.md to categorize potential threats and prioritize mitigation efforts.

The Three-Tier Risk Taxonomy

According to the source code in pm-execution/skills/pre-mortem/SKILL.md (lines 27-43), the pre-mortem skill organizes identified risks into three distinct psychological and severity-based categories. This taxonomy helps teams distinguish between substantiated threats and cognitive biases.

Tigers: Real Launch Threats

Tigers represent substantive, validated problems that could realistically derail the product launch. These are high-impact risks with genuine probability, requiring active mitigation strategies. In the classification workflow, Tigers receive the highest attention and are subjected to the secondary urgency classification described below.

Paper Tigers: Overblown Concerns

Paper Tigers are concerns that appear serious on the surface but are objectively unlikely to occur or are disproportionate to their actual impact. The skill categorizes these separately to prevent teams from over-investing resources in hypothetical scenarios while still acknowledging the anxiety they generate.

Elephants: Hidden Team Worries

Elephants refer to hidden or unspoken worries that team members are reluctant to discuss openly. The pre-mortem skill explicitly surfaces these to prevent organizational silence from blindsiding the launch, ensuring that tacit knowledge and unvoiced fears receive visibility in the risk assessment.

Secondary Classification: Tiger Urgency Levels

Once a risk is classified as a Tiger, the pre-mortem skill applies a second-level urgency classification defined in pm-execution/skills/pre-mortem/SKILL.md (lines 44-53). This tier determines the temporal priority of mitigation efforts based on launch impact.

The three urgency categories are:

  • Launch-Blocking – Critical issues that must be resolved before the product can go live. These represent hard gates that prevent release.
  • Fast-Follow – Significant risks that can be addressed within the first 30 days post-launch without catastrophic impact, but require rapid remediation.
  • Track – Risks that are merely monitored; active remediation occurs only if the issue materializes post-launch.

The /pre-mortem command explicitly references this taxonomy when generating the risk table, as implemented in pm-execution/commands/pre-mortem.md (lines 38-42).

How to Invoke the Classification System

To classify risks using the pre-mortem skill, invoke the slash command with your launch context. The skill automatically executes the two-step classification workflow.

/pre-mortem We're launching a self-serve billing portal next month

The workflow executes the following steps:

  1. Identify Risks – Analyzes the provided PRD or plan to surface potential failure points
  2. Classify Risks – Assigns each risk to Tiger, Paper Tiger, or Elephant based on the criteria in lines 27-43
  3. Tag Tigers – Marks every Tiger as Launch-Blocking, Fast-Follow, or Track according to severity and time-to-mitigate
  4. Output – Generates a structured markdown report with separate sections for each risk type and a prioritized table for launch-blocking Tigers

Sample Output Structure

The generated report follows a strict markdown format that reflects the classification hierarchy:


## Pre-Mortem: Self-Serve Billing Portal

### Tigers (Real Risks)

| # | Risk | Severity | Category |

|---|------|----------|----------|
| 1 | Payment gateway integration fails under load | Launch-Blocking | Technical |
| 2 | Customers cannot find the pricing page | Fast-Follow | User |
| 3 | Billing data not GDPR-compliant | Launch-Blocking | Operational |

### Paper Tigers (Overblown Concerns)

- "The UI looks too modern for enterprise users" – branding team worries, but no market data supports it.

### Elephants (Unspoken Worries)

- No clear ownership for post-launch support of the new billing API.

### Action Plans for Launch-Blocking Tigers

| # | Risk | Mitigation | Owner | Due |

|---|------|------------|-------|-----|
| 1 | Payment gateway integration fails under load | Load-test with 2× expected traffic, add circuit-breaker | Eng-Lead | 2026-07-15 |
| 3 | Billing data not GDPR-compliant | Conduct data-privacy audit, update consent flow | PM-Ops | 2026-07-10 |

Implementation in Source Code

The classification logic resides in two primary files within the phuryn/pm-skills repository:

This architecture separates the taxonomic definitions from the command execution layer, allowing the classification system to remain consistent across different invocation contexts.

Summary

  • The pre-mortem skill uses a two-level classification system: primary risk types and secondary urgency tags for Tigers.
  • Three primary categories exist: Tigers (real threats), Paper Tigers (overblown concerns), and Elephants (hidden worries), defined in SKILL.md lines 27-43.
  • Tigers receive three urgency classifications: Launch-Blocking (pre-launch fix required), Fast-Follow (30-day post-launch window), and Track (monitor only), defined in lines 44-53.
  • The /pre-mortem command in pre-mortem.md (lines 38-42) automates this classification and outputs structured markdown reports.
  • This system prevents resource waste on Paper Tigers while ensuring Elephants receive visibility and Tigers receive appropriate temporal prioritization.

Frequently Asked Questions

What distinguishes a Tiger from a Paper Tiger in the pre-mortem skill?

A Tiger is a substantiated risk with real probability of derailing the launch, while a Paper Tiger appears serious but is objectively unlikely or overblown. The distinction appears in pm-execution/skills/pre-mortem/SKILL.md lines 27-43, where Tigers require active mitigation and Paper Tigers require acknowledgment without resource allocation.

How does the pre-mortem skill determine if a Tiger is Launch-Blocking versus Fast-Follow?

The classification depends on the time-to-mitigate and impact severity. Launch-Blocking Tigers (lines 44-53 of SKILL.md) prevent release entirely until resolved, while Fast-Follow Tigers can tolerate a 30-day post-launch remediation window without catastrophic failure. The skill evaluates whether the issue can be fixed after launch or constitutes a hard gate.

What are Elephants in the context of pre-mortem risk classification?

Elephants represent hidden or unspoken worries that team members are reluctant to voice, defined in the taxonomy section (lines 27-43). The pre-mortem skill explicitly creates space for these concerns to prevent organizational silence from blindsiding the launch, distinguishing them from overt risks (Tigers) and exaggerated fears (Paper Tigers).

Where is the risk classification logic implemented in the pm-skills repository?

The classification taxonomy is defined in pm-execution/skills/pre-mortem/SKILL.md (lines 27-53), while the command interface that executes the classification resides in pm-execution/commands/pre-mortem.md (lines 38-42). The skill markdown contains the categorical definitions, and the command markdown implements the user-facing invocation and output formatting.

Have a question about this repo?

These articles cover the highlights, but your codebase questions are specific. Give your agent direct access to the source. Share this with your agent to get started:

Share the following with your agent to get started:
curl -s "https://instagit.com/install.md"

Works with
Claude Codex Cursor VS Code OpenClaw Any MCP Client

Maintain an open-source project? Get it listed too →