Guide · Agent skill
How to use the Decision Post-Mortem agent skill (post-mortem)
Updated · written with AI from the project's own documentation and checked against our catalogue data
The Decision Post-Mortem skill (post-mortem) is an agent skill for Claude that reviews a business decision after it has played out. It compares actual results with the success and kill criteria written before the decision, revisits concerns that were raised and overruled, and checks the original assumptions. This guide explains what the skill does, how to install it, how to trigger it, and what it needs to work well.
Decision Post-Mortemalirezarezvani/claude-skills · MIT Download the skill Quick start
- 1Download and unzip. Download the skill ZIP from this page and unzip it to get SKILL.md.
- 2Choose a location. Place the folder in ~/.claude/skills/ for all projects, or in a project's .claude/skills/ folder for one project. In the Claude apps, upload it under Settings → Capabilities → Skills instead.
- 3Confirm companion files. Check that the decision-logger, cs-chief-of-staff and executive-mentor skills are present in the repository if you want the linked routes to work.
- 4Gather the inputs. Collect the decision record from /c-level-agents:decide, the execution plan from /c-level-agents:execute, and your actual outcome data.
- 5Run the command. Run /c-level-agents:post-mortem with the decision name, or ask for a post-mortem in natural language at the 90-day checkpoint or when a kill criterion triggers.
- 6Review the record. Check the scoring tables, dissent and assumption sections, then act on the status and forward actions.
What is the post-mortem skill and what does it make the agent do?
An agent skill is a folder of instructions that Claude loads when a request matches the skill's purpose. This skill is part of the c-level-agents toolkit in the alirezarezvani/claude-skills repository, and it is invoked as /c-level-agents:post-mortem [decision].
Its job is an honest retrospective on a decision that has already been executed. The agent scores results against the criteria recorded before the decision, not criteria written afterwards. It then produces a structured post-mortem record with a WIN, PARTIAL, LOSS or MIXED status and suggested forward actions.
- Scores success criteria and kill criteria against actual outcomes, marking each as met or triggered
- Revisits preserved dissent from the earlier boardroom discussion and asks whether each concern materialised
- Audits the original brief's assumptions as held, not held, or partly held
- Records what went right, what went wrong, and process lessons
- Saves the record to ~/.claude/postmortems/ as a Markdown file named by date and slug
Install the post-mortem skill from the ZIP or the Claude apps
Download the skill ZIP from this page and unzip it. The skill is a single file, SKILL.md. Place the unzipped folder in one of two locations, depending on how you want to use it.
If you use Claude Code and want the skill available across all your projects, put it in ~/.claude/skills/. To limit it to one project, put it in that project's .claude/skills/ folder.
If you use the Claude apps, you can upload the skill instead. Go to Settings, then Capabilities, then Skills, and upload it there.
- Personal, all projects: ~/.claude/skills/
- Single project: .claude/skills/ inside the project folder
- Claude apps: Settings → Capabilities → Skills
Prepare the inputs before you run it
The skill scores results against criteria that already exist. Before running it, you need the decision record from /c-level-agents:decide, the execution plan from /c-level-agents:execute, and the actual outcome data you want to score. The skill does not collect metrics for you.
The source also says the companion skills decision-logger, cs-chief-of-staff and executive-mentor must be present in the repository for the linked routes to work. Without them, the cross-references will not resolve.
- Decision record from /c-level-agents:decide
- Execution plan from /c-level-agents:execute
- Actual outcomes: metrics, events and customer signals you supply
- Access to the local ~/.claude folder so the record can be saved
How do you trigger the skill with natural requests?
You can call the command directly with the decision name, as in /c-level-agents:post-mortem [decision]. The skill's description also says when it should run, so a natural request that matches those situations should bring it in as well.
The skill is meant for three moments: when a decision reaches its 90-day review checkpoint, when a kill criterion triggers, and when a major decision is reversed. It also suggests a quarterly pass over all decisions from the past quarter.
- "It's been 90 days since we changed our pricing. Run a post-mortem on that decision against the success metrics we set."
- "Our kill criterion for the EU expansion just triggered. Score that decision with the post-mortem skill."
- "We reversed the hiring freeze. Do a post-mortem and check whether the original dissent was right."
- "Review all decisions from last quarter against their pre-committed criteria."
/c-level-agents:post-mortem [decision]Read the output: scoring, dissent and assumptions
The record follows a fixed structure. The outcome scoring tables compare each success criterion and kill criterion with its threshold and actual value, then mark it as met or triggered. The overall status is one of WIN, PARTIAL, LOSS or MIXED.
The dissent section restates each original concern and marks whether it materialised, with a cost if it did and a one-sentence lesson. The assumption audit does the same for the brief's assumptions. The process lessons section asks whether isolation worked, whether devil's advocate concerns played out, and whether the cadence was right.
The Status section sets the next step. A WIN is archived with its lesson logged. A LOSS schedules a follow-up boardroom through /c-level-agents:brief.
- Outcome Scoring: success and kill criteria tables
- What We Got Right and What We Got Wrong
- Preserved Dissent — Revisited
- Assumption Audit
- Process Lessons
- Forward Actions as a checklist
Safety notes: scripts and network
According to the skill's listing, it does not run scripts and does not use the network. Its work is reading the records and outcome data you give it and writing the post-mortem file to ~/.claude/postmortems/.
Because the scoring depends entirely on the metrics you supply, check those numbers before you run the skill. The skill does not verify their accuracy, so do not rely on it for verified financial or legal conclusions.
- No scripts are run by this skill
- No network use is listed
- Output is saved locally under ~/.claude/postmortems/
- Check the accuracy of your input metrics yourself
Tips for useful post-mortems
The skill works best when the criteria were written down before the decision was made. If they were not, the skill can only offer after-the-fact judgement, and it warns against retro-fitting criteria. Recording thresholds at decision time is the single most useful habit.
Revisiting dissent builds a record over time. The source notes that dissenters were often directionally right, so keep the dissent column from earlier boardroom memos available for the review.
- Write success and kill criteria with thresholds before acting
- Gather actual metrics for each criterion before running the review
- Keep boardroom dissent notes so they can be scored later
- Run it at the 90-day mark rather than only after a loss
What people use it for
- Scoring last quarter's pricing change against its pre-committed success metrics
- Checking whether a kill criterion was triggered on a market entry decision
- Testing whether a dissenting view from a past boardroom discussion turned out to be right
- Running a quarterly review of all decisions from the past three months
- Logging process lessons about how decisions were isolated, challenged and paced
Frequently asked questions
How do I install the post-mortem agent skill?
Download the skill ZIP from this page and unzip it. Put the folder in ~/.claude/skills/ to use it across projects, or in a project's .claude/skills/ folder for a single project. In the Claude apps, upload it under Settings → Capabilities → Skills.
What does the post-mortem skill need to run?
It needs the decision record from /c-level-agents:decide, the execution plan from /c-level-agents:execute, and actual outcome data you supply. It also needs access to ~/.claude to save the record, and the related skills decision-logger, cs-chief-of-staff and executive-mentor must be present for linked routes to work.
Is the post-mortem skill free to use?
The source does not state price or licence terms beyond the listing's MIT licence, so check the repository for the licence text before using it.
Does the skill run scripts or use the internet?
According to the listing, it runs no scripts and uses no network. It reads the records and outcome data you provide and saves the post-mortem file locally.
Can I use it if I did not write criteria before the decision?
The skill is designed around criteria written before the decision. Without them it can only give after-the-fact judgement, and it warns against retro-fitted criteria. Using it is still possible, but the scoring will be weaker.
How is this different from the executive-mentor post-mortem?
The source describes /em:postmortem in executive-mentor as an adversarial single-decision post-mortem. This skill instead scores a decision against its pre-committed criteria and revisits preserved dissent as part of the c-level-agents sprint loop.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use this skill if you already record decisions with written success metrics and want a disciplined review at the 90-day mark or when a kill criterion triggers. Pick something else if you have no prior decision records, or if you need verified financial or legal conclusions, since the scoring relies entirely on the metrics you supply.
Download the skill