Transparency X-ray

Exactly what left your machine

Every number on this page came from structural metadata alone — counts, timestamps and salted hashes. Never your code, never your prompts, never a file name. Below: the full trail, from the raw signals to every single point of the score.

← @bia-codes22 C- 35.1 last 30 days · 0 sessions · 0 active days algorithm v3

The privacy contract

Never leaves your machine

  • Your code and your diffs
  • Your prompts and any conversation text
  • File paths and file names
  • Repository and project names — only salted hashes travel
  • Secrets, URLs and environment values

Everything that is sent

  • Structural counts per session — agents dispatched, files touched, tool calls
  • Session start and end timestamps
  • Salted, irreversible project and session hashes
  • Tool names only (e.g. "Bash", "mcp__gmail") — never their inputs or outputs

The fleet, in numbers

0 agents orchestrated The real fleet: every subagent you dispatched across 0 workflows in the window.
0 workflows invoked Multi-agent workflows you started.
0 peak concurrent Hardware-capped at min(16, cores − 2). A ceiling of your CPU — not the size of your fleet.
0 sessions Across 0 active days in the window.

How your score was computed

All six pillars, every sub-term in the open: the real inputs, the formula, and the points it produced. Hill(x, k, p) is a saturation curve — quick to rise, then it flattens — so volume can never buy the top of a pillar.

Orchestration

11.3 / 30 pts
  • Workflow density

    Every multi-agent workflow you invoked — the curve saturates around 10.

    W = 0 11 × Hill(W, 10, 1.2)
    +0.0 / 11
  • The fleet (delegation)

    The total of agents you dispatched — the real fleet, and the dominant term of this pillar.

    G = 0 12 × Hill(G, 120, 1)
    +0.0 / 12
  • Concurrent peak (bonus)

    Your peak of agents running at the same instant — a hardware-capped bonus, no longer the main axis.

    Fmax is capped by your CPU at min(16, cores − 2) — it measures the concurrent peak, not your fleet. The fleet is G = 0.

    Fmax = 0 4 × Hill(Fmax, 3, 2)
    +0.0 / 4
  • Breadth of fronts

    Distinct projects you orchestrated across — conducting several fronts is part of the craft.

    proj = 0 3 × Hill(projects, 4, 1)
    +0.0 / 3

Verification

6.6 / 18 pts
  • Verify after edit

    The share of editing sessions where you ran tests, builds or re-reads afterwards.

    r = 0% ne = 0 8 × r^0.7
    +0.0 / 8
  • Verification breadth

    How many distinct ways you verify — tests, builds, lint, browser — out of 4.

    mod = 0 6 × (mod ÷ 4)
    +0.0 / 6
  • Human sign-off

    AskUser moments plus explicit plans — checkpoints where a person approves before the fleet moves.

    S = 0 4 × Hill(S, 8, 1)
    +0.0 / 4

Tooling

8.5 / 16 pts
  • Tool diversity

    Distinct tool names used at least twice in the window.

    T = 0 6 × Hill(T, 12, 1.5)
    +0.0 / 6
  • Advanced tooling

    ToolSearch and Skill invocations — reaching beyond the defaults.

    A = 0 6 × Hill(A, 10, 1)
    +0.0 / 6
  • Ecosystem (MCP + plugins)

    Distinct MCP servers reached, derived server-side from the mcp__ tool names.

    E = 0 4 × Hill(E, 4, 1)
    +0.0 / 4

Delivery

1.3 / 14 pts
  • Pull requests

    PRs linked in the window — shipped work, not just motion.

    P = 0 8 × Hill(P, 6, 1)
    +0.0 / 8
  • Magnitude

    Distinct files touched across the window.

    M = 0 4 × Hill(M, 40, 1)
    +0.0 / 4
  • Efficiency

    PRs per unit of orchestration — η = P ÷ (1 + W + G/5).

    η = 0.00 2 × Hill(η, 0.5, 1)
    +0.0 / 2

Craft & Discipline

7.1 / 12 pts

Craft starts at the full 12 pts and can only lose points — four capped penalties for messy runs.

  • Compaction penalty

    Average context compactions per session — long, bloated contexts cost points.

    comp = 0.00 − min(comp × 2.5, 2.5)
    +0.0 cap 2.5
  • API error penalty

    Average API errors per session.

    err = 0.00 − min(err, 2.0)
    +0.0 cap 2
  • Denial penalty

    Average tool denials per session — permissions friction.

    den = 0.00 − min(den × 1.5, 1.5)
    +0.0 cap 1.5
  • Interruption penalty

    Average interruptions per session — how often a run had to be stopped.

    int = 0.00 − min(int, 2.0)
    +0.0 cap 2

Autonomy

0.8 / 10 pts
  • Loops

    ScheduleWakeup and Monitor events — work that keeps running without you.

    L = 0 5 × Hill(L, 6, 1)
    +0.0 / 5
  • AFK leverage

    Your longest run with you away — it only counts with real verification behind it.

    Gate: only scores when Verification is above 4 pts. AFK without verification is not autonomy.

    Rafk = 0 3 × Hill(Rafk, 40, 1) × gate
    +0.0 / 3
  • Consistency

    Active days across the month, discounted by how concentrated the activity is (Gini).

    days = 0 Gini = 0.00 2 × (days ÷ 30) × (1 − 0.5 × Gini)
    +0.0 / 2

Every aggregate signal

The labelled totals behind the formulas — the same numbers you can check against the raw JSON below.

0 sessions
0 active days
0 workflows invoked
0 agents orchestrated
0 peak concurrent
0 files touched
0 PRs linked
0 advanced tool uses
0 AskUser checkpoints
0 explicit plans
0 scheduled wakeups
0 monitors
0 longest AFK run
0 compactions
0 API errors
0 distinct tools
0 MCP servers derived server-side
0 distinct projects
0 ecosystem uses (server-derived) derived server-side

Per-session metadata (a sample)

The exact per-session structure inside the payload: counts, timestamps and salted hashes. A hash cannot be walked back to a repo, a path or a file name.

No sessions in the current 30-day window.

Don't trust — verify

Download the exact raw payload this page was rendered from: the complete record of everything that ever left your machine, in one file you can read in a minute.

Download the raw JSON

AgenticLevel reads structure, not content. If any number here surprises you, the raw JSON above is the ground truth.