Perfolio is developed out in the open. Below is everything that's live today, in reverse order — followed by the themes we're working on now, next, and later.
In reverse order. This section is what exists today.
pfl mcp install now wires Perfolio into Cursor, Claude Desktop, Codex, Google Antigravity, VS Code, Windsurf, and Gemini CLI — each written in its own native config schema, chosen from an interactive multi-select. Claude Desktop is bridged through mcp-remote automatically, and every write is non-destructive with a backup.
Log a win straight from your terminal in seconds. pfl log sends it to your review inbox with lightweight significance coaching so only real signal lands; pfl auth stores your token in the OS keychain. Zero runtime dependencies, published to npm.
Companies rolling Perfolio out to their teams can connect verified-domain SAML single sign-on for secure, seamless access for everyone in the org.
Configured sensitive fields are encrypted with AES-256-GCM using per-user data keys wrapped by Google Cloud KMS, shipped alongside a broad audit-hardening pass. Supabase Row-Level Security isolates every user's records.
A clearer live view of how your evidence covers each ladder competency, with a compact cycle tracker so you always know where the gaps are before your review window locks.
Any MCP-capable agent can connect to Perfolio over HTTPS with a scoped, revocable token. Agents read your ladder coverage and evidence, and can propose new entries — always behind a human review gate before anything touches your record.
A full guided product tour that lets you explore capture, calibration, and the promotion packet without creating an account.
Read-only GitHub, Slack, and Gmail connectors turn everyday work into evidence; ladder calibration maps each item to your rubric; the Sponsor Warmth Radar keeps advocates warm; and one export produces a citation-backed dossier in Markdown or PDF. Your record is fully portable — export everything as Markdown, JSON, or CSV, anytime.
Directions, not promises — deliberately undated.
Broader AI-agent coverage
Expanding the MCP toolset and testing across more agents so capture works wherever you code.
Faster, cleaner capture
Tightening the review inbox and significance signals so accepting a real win takes seconds.
Beyond engineering ladders
The same evidence-against-a-rubric problem exists outside engineering. Bringing more role families and rubric templates.
Scheduled background capture
Opt-in, hands-off capture that keeps your record current between review cycles.
Sharper manager & coaching surfaces
Better inputs for 1:1s and growth conversations — without ever crossing the employee-owned boundary.
Deeper integrations
More read-only sources for richer, verifiable context around your work.
Team-level insight (aggregate only)
Cohort adoption and coverage trends for orgs — never private entries or individual scores.
Your ideas
The biggest driver of this column is what you ask for. Requests shape priorities.
We don't attach dates because priorities shift as we learn from you. Items move left as they ship — and occasionally get cut. Only what's under Shipped is guaranteed to exist today.
Have something you want to see here? User requests shape this roadmap more than anything else.