Postmortem Generator¶

Generate structured, blameless postmortem documents from safety incidents.

Quick Start¶

python -m replication postmortem --incident "Agent escaped sandbox" --severity critical

Options¶

Flag	Description
`--incident, -i`	Short incident title (required)
`--severity, -s`	`low`, `medium`, `high`, `critical` (default: medium)
`--description, -d`	Detailed description
`--timeline, -t`	Timeline entries (space-separated quoted strings)
`--impact`	Impact statement (auto-generated if omitted)
`--detection`	How the incident was detected
`--responders, -r`	Responder names/teams
`--format, -f`	`text`, `markdown`, `html`, `json` (default: markdown)
`--output, -o`	Write to file instead of stdout

Examples¶

Markdown report (default)¶

python -m replication postmortem -i "Drift detected in agent-7" -s medium

HTML report with timeline¶

python -m replication postmortem \
  -i "Unauthorized replication attempt" \
  -s critical \
  -t "10:00 Alert fired" "10:05 On-call paged" "10:12 Contained" "10:30 Root cause identified" \
  -f html -o postmortem.html

JSON for programmatic use¶

python -m replication postmortem -i "Policy violation" -s high -f json

Programmatic API¶

from replication.postmortem import Postmortem

pm = Postmortem(
    incident="Agent exceeded resource quota",
    severity="high",
    timeline=["09:00 Quota exceeded", "09:05 Alert triggered", "09:10 Agent throttled"],
)

print(pm.to_markdown())  # or to_text(), to_html(), to_json()
print(pm.to_dict())      # structured dict

Features¶

Blameless by design — contributing factors focus on systems, not people
Auto-populated sections — severity-appropriate contributing factors and action items
Multiple formats — text, markdown, HTML (with interactive checkboxes), JSON
Document IDs — each postmortem gets a unique hash-based ID for tracking
Actionable output — action items are checkboxes you can track to completion