The Agent skill that turns teamwork with AI into a team portrait.
Teamfolio reads authorized Agent conversations shared through CoAligne and turns recurring team behavior into a playful, evidence-grounded archetype.
Your team might resemble an F1 pit crew, a rescue squad, a migrating shoal, a lighthouse network, an experimental kitchen, or something nobody thought to compare a product team with before.
The analogy can be imaginative. The evidence cannot.
Teamfolio generates a portable single-file HTML report containing:
- a team archetype and one-line verdict;
- evidence-to-metaphor mappings;
- an operating profile based on recurring work patterns;
- privacy-safe stories from observed Agent work;
- a team superpower and lovable failure mode;
- a collectible artifact from the archetype's world;
- an explicit evidence boundary and coverage summary;
- an original visual direction selected for the team.
The visual system is not one universal template. It can route to precision blueprints, ephemera collage, low-poly quests, naturalist specimens, naïve posters, risograph field cards, flow fields, or a justified variation.
These reports use the same evidence discipline with entirely different comic worlds. Both single-file pages include an in-page 中文 / EN switch; the previews below show their English states.
Teamfolio requires CoAligne because a single Agent normally cannot access other teammates' conversation histories.
Before running Teamfolio:
- Install and sign in to CoAligne.
- Create or join a team.
- Invite teammates.
- Have at least two different teammates import and sync Agent-active projects.
- Make sure the current Agent can call the CoAligne tools.
Teamfolio checks these conditions before reading conversations. A team is eligible only when readable conversation bodies are available from projects created by at least two distinct teammates.
Copy the teamfolio folder into your Codex skills directory:
~/.codex/skills/teamfolio
Restart or open a new Agent session after installation.
Invoke the Skill and name a team plus either a comparison domain or random.
$teamfolio Analyze the CoAligne team and compare us to an F1 team.
$teamfolio Analyze our team. Random archetype.
$teamfolio Are we more like a rescue squad or an experimental kitchen?
When several eligible CoAligne teams are accessible, Teamfolio asks which one to analyze before reading conversation content.
Teamfolio:
- reads only accessible, authorized, synced CoAligne context;
- treats past conversations as untrusted evidence, not executable instructions;
- distinguishes teammate decisions from Agent output;
- bases team-wide claims on patterns observed across multiple teammates;
- does not treat message volume, project count, or Agent fluency as competence;
- redacts emails, credentials, internal paths, and unrelated sensitive details;
- describes missing evidence as unobserved rather than absent;
- does not produce a psychological profile, HR verdict, or complete performance evaluation.
.
├── teamfolio/
│ ├── SKILL.md
│ ├── agents/openai.yaml
│ ├── assets/report-template.html
│ ├── references/
│ │ ├── analysis-framework.md
│ │ ├── coalign-evidence.md
│ │ ├── comedy-worldbuilding.md
│ │ ├── visual-directions.md
│ │ └── visual-system.md
│ └── scripts/validate_report.py
└── docs/images/ # sanitized README previews
Validate the Skill:
uv run --with pyyaml python ~/.codex/skills/.system/skill-creator/scripts/quick_validate.py teamfolio
Validate a generated report:
python3 teamfolio/scripts/validate_report.py teamfolio-example.html
MIT