c16
c16 / wf5.m2 - Agent Control Plane — hi-fi unconstrained (wf5.m2) / Agent Profile·control-plane-v1/agent-profile·draft
Agent Inventory agent-as-actor · one agent, all the way down
100/ 100 · target ≥ 80Standard

Claude Code

governedinstrumented

Anthropic · cli · claude-opus-4-8 · pid:865:claude-code

Claude Code is Standard (100/100): no deductions. Credential risk Lower — revocable OAuth, enterprise-backed. Governed.

cost

$214.50

tokens

6.9M

cred risk

Lower

success

93%

data-risk

12

Worst observed

100

no prior regression

Why this score — deduction ledger

100 base − 0 deducted
100 kept

No deductions — clean posture, full 100 kept.

Credential risk Lower (revocable OAuth, enterprise-backed). Policy-bound.

Identity & Access

user
mark
email
mark@qpoint.io
org_id
org_qpoint
auth_method
oauth_jwt
cred_source
idp_oauth
cred_risk
Lower

revocable OAuth, enterprise-backed

Process & Runtime

name / version
Claude Code · v2.1.161
model
claude-opus-4-8
binary_path
/Users/mark/.local/bin/claude
cwd
/Users/mark/web/qpoint/bob-wire
config_source
/Users/mark/.claude/settings.json
first → last
2026-05-08 → 2026-06-03

Token & Cost breakdown

6.9M tokens · 1.2k calls · $214.50
6.9Mtokens
input17%1.2M
output3%180.0k
cache read70%4.8M
cache write9%600.0k
reasoning1%90.0k

By model

claude-opus-4-86.9M tok · 1.2k calls · $214.50
See spend governance & caps →

Is it working? — task outcomes

measured

Of every 100 tasks

93 clean
11
success intervene rework

93%

success

88

quality

142

tasks

From run-outcome telemetry.Fleet effectiveness →

What it did — recent activity

1.8k observed events
tool
1.2k
file
410
mcp
124

320

reads

45

writes

248

mcp fp

12

data-risk

Open data & capability findings →

Sensitive data touched

data-risk 12
Path
Category
Severity
Opens
./.envsecret_filehigh3
~/.aws/credentialscloud_credentialshigh1

MCP servers

footprint 248 · 2 × 124 calls

mcp.figma.com

84

github-mcp

40

Network Egress

detail-level signal · not an organizing surface
Host : Port
TLS
Allowlisted
api.anthropic.com:443TLSv1_3yes
mcp.figma.com:443TLSv1_3no

Do next

Proposed components: AgentPostureHeader · ScoreLedger · TokenCostBreakdown · OutcomeSplit · EventFamilyBars. Derived: deductions (factors × FACTOR_PTS) · delta vs worst_score · data-risk (Σ severity×opens) · MCP footprint · outcome split · activity slice · isCompoundRisk — from app/data/c16-fleet.ts (qcontrol /api shape).