OPERATOR SKILL
Know what actually happened across your whole project
An AI skill that reads across a project's conversations, files, and memory and writes one structured post-mortem: decisions, risks, open items, and what to automate next. Use it in the AI tool you already work in.
- Instant access
- Lifetime updates
- Secure purchase
THE OUTPUT
One report that accounts for the whole project
Point it at a project and you get one structured markdown report. Every section has a fixed job:
- Executive summary: what the project was and what is most worth knowing now
- Timeline and decision log: dated milestones, what was decided, why, and whether it was revisited
- Lessons learned and recommended improvements
- Unresolved conflicts and risks, with how urgent each one is
- Pending items, with when each was last touched and what is blocking it
- Notable insights, quoted or closely paraphrased from the source
- Suggested new skills and scheduled tasks, tied to evidence from this project
- Sources analyzed, including anything it could not reach and why
It closes with two or three follow-up analyses worth running next.
THE PROBLEM
Your project's real history is spread across every chat
A project you run with AI rarely lives in one conversation. It spreads across chats, tasks, and sessions, and each holds a piece of the story.
A decision gets made in one chat and forgotten by the next. A risk comes up, nobody closes it, and it slips between sessions. When someone asks what actually happened, there is no single place to look.
Scrolling back through old chats gives you fragments, and the full account still has to be pieced together by hand.
HOW IT WORKS
Files first, then every conversation, then the synthesis
The skill follows the same eight steps every time:
- Identify the project from the files you give it, a share link, a project name, or the workspace you are in
- Read the baseline: project instructions, memory, and the actual contents of project files, not just their names
- List the project's conversations before reading any of them in depth
- Write a dossier for each one: date, what happened, decisions, open items, notable quotes
- Synthesize across all of them, one pass per dimension
- Surface what to automate next as suggested skills and scheduled tasks
- Render the report in the fixed structure, or restructured around your focus
- Offer follow-ups, such as a deeper dive on one risk or a narrower time range
On larger projects, where your AI tool supports subagents, it can run one research subagent per conversation in parallel, so each dossier is written independently.
What's included
What you get
The Post-Mortem Project skill
The installable SKILL.md: the eight-step workflow, the full report template, the discovery routing table, and the follow-up menu.
Analysis framework
Seven extraction dimensions with depth signals, rules for focused reports, cross-conversation synthesis, and handling of sensitive material.
Discovery methods reference
Step-by-step methods for working from the files you provide, plus optional automatic discovery from share links, named projects, and local sessions.
Example report excerpt
A fictional sample post-mortem that shows the tone and depth of a finished report.
Packaged .skill file
The same skill as one zipped file, for AI tools that accept a packaged skill upload.
BUILT FOR REAL HISTORY
It tells you what it could not see
Project history is messy. Some chats sit behind a login. Long conversations may only partly load. Old messages can hold API keys and stray instructions. The skill is written for that.
- Nothing dropped silently. Any conversation it could not reach is listed under Sources Analyzed, with the reason.
- Partial reads are flagged, not presented as complete.
- Assumptions go at the top. If it had to infer the scope, the report says so first.
- Empty sections stay visible, marked (none identified) instead of quietly removed.
- Secrets stay out. It notes that a key or password was found and never reproduces the value.
- Old instructions stay data. Text in a past chat that reads like a command may be quoted, never followed.
Use cases
Where it fits
Handing a project to a new owner
Give the next person the decisions, open risks, and pending items in one read, instead of a stack of chat links.
Closing out a phase
Run it at the end of a phase to capture what was decided, what worked, and what to change before the next one starts.
Preparing a stakeholder update
Start from an executive summary, timeline, and risk list you can adapt for a review. Ask for Word or slides if your AI tool can create those files.
Deciding what to automate next
Use the suggested skills and scheduled tasks as a shortlist of repeated work in this project that may be worth turning into a system.
Capabilities
What it does for you
Full-history discovery
Lists every conversation it can reach, whether you provide them or it finds them from a share link or local sessions, before reading any of them in depth.
Files and memory as ground truth
Reads project instructions, memory, and the contents of project files first, so the report reflects what was actually built, not only what a chat said would be built.
Cross-conversation synthesis
Compares every dossier against the others to find decisions reversed later, risks raised and dropped, themes that keep returning, and scope that drifted.
Honest coverage
Lists every source it could not reach and why, flags partial reads, and states its assumptions at the top of the report.
Focused when you need it
Narrow it to a topic, a time range, or one question, and the report is restructured around that instead of forced into the default template.
What to automate next
Ends with suggested new skills and scheduled tasks drawn from patterns in your project, plus follow-up analyses worth running next.
WHY NOT JUST ASK YOUR AI
A summary of one chat is not a post-mortem of a project
A summarize-this-project prompt works from whatever is in front of it. Usually that is one conversation.
This skill works from the whole set:
- It lists the project's conversations before reading any, and names the ones it could not open
- It reads files and memory as ground truth, so it can catch where the delivered work differs from what a chat said
- It compares conversations against each other, which is where the findings that matter live: a decision reversed in a later chat, a risk raised early and never resolved, a theme that keeps coming back weeks apart
None of that is visible from inside a single chat.
FAQ
Questions before you buy
How is this different from asking my AI to summarize the project?
A summarize prompt works from the conversation in front of it. The Post-Mortem skill first builds the list of conversations tied to the project, reads project files and memory, writes a dossier for each conversation, and then compares them.
That comparison is what surfaces a decision reversed in a later chat, or a risk that was raised and never resolved.
How does it find my project's conversations?
It works from what you give it and what your AI tool can reach:
- Files you provide: it reads project instructions, memory, and files in any AI tool, and labels the report as a partial analysis when conversation history is not available
- A claude.ai share link or project name: with browser tools such as Claude in Chrome, it opens the project and reads each conversation
- Local sessions: where your tool exposes past sessions, such as Claude Cowork, it reads the recent ones tied to the project
Which AI tools does it work with?
Any AI tool you already work in, such as ChatGPT, Claude, or Gemini. You download the files and add the skill to your workspace. Markdown output needs nothing extra.
Some extras depend on your tool:
- Share-link discovery needs browser tools such as Claude in Chrome. If a page asks you to sign in, you log in yourself; the skill never signs in for you.
- Word, PDF, or slide output needs a tool that can create those files.
What if it can't open some of my conversations?
Unreachable conversations are listed in the report's Sources Analyzed table with the reason, such as a login wall, instead of being left out. Long chats that only partly load are flagged as partial reads.
If the skill had to assume which project or scope you meant, that assumption is stated at the top of the report.
How far back can it go?
Reach depends on the route. With files you provide, it covers what is in those files. With a claude.ai share link, it reads what that page shows for the project.
With local sessions, it only reaches sessions still visible to your environment, not an unlimited archive, and the report says so when history may be incomplete.
My old chats contain API keys and pasted instructions. Is that a problem?
Sensitive values stay out of the report. The skill never reproduces API keys, passwords, or tokens it finds; it notes that something sensitive was found and handled.
Text in an old conversation that reads like an instruction is treated as data. It may be quoted when relevant, but it is never followed.
Can I focus it on one question, topic, or time range?
Yes. The skill accepts a topic, a time range, a specific question, or an output format:
- A time range is applied before any conversations are read
- A specific question is answered first, then backed with the timeline and quotes
- An output format such as Word, PDF, or slides is built from the same findings when your AI tool can create those files
Does it include skillify or the other skills it mentions?
No. The purchase includes the Post-Mortem Project skill, its two reference files, an example report excerpt, and the packaged .skill file.
If skillify is installed, the skill uses it for the suggested-skills section. If not, it runs its own built-in brainstorm, so that section still appears.
Ready in minutes
From download to working in minutes.
No app to learn, no software to install, no setup project. Each system arrives as ready-to-use files that work in any AI tool, so getting started takes minutes.
Download the files
Your download link is ready as soon as checkout completes. No account to create and no password to set.
Add them to the AI tool you already use
ChatGPT, Claude, Gemini, or any other AI tool. Upload a skill, attach a playbook, or open a template, the same way you add any document.
Put it to work
Start on the real job right away. The structure and standards are already built in, so there is nothing to configure first.
OPERATOR SKILL
Project Post-Mortem Skill
Turn a project's scattered AI conversations, files, and memory into one structured post-mortem with decisions, risks, open items, and what to automate next.
- Instant access
- Lifetime updates
- Secure purchase