Agent Ledger

Track AI agent earnings, tasks, and payments. Use when logging completed work, checking balance, viewing payment history, or generating invoices. Supports cr...

Security Scan

The skill is a small, file-based ledger (bash + jq) that stores data under ~/.agent-ledger and does not request credentials or make network calls; its requirements and behavior match its description.

Purpose & Capability

Name/description (agent earnings ledger) align with required binaries (bash, jq), storage in ~/.agent-ledger, and the provided CLI script. Nothing requested is extraneous to bookkeeping.

Instruction Scope

SKILL.md instructs running scripts/ledger.sh and storing data in ~/.agent-ledger (override via AGENT_LEDGER_DIR) — appropriate for the purpose. Minor implementation issues: user-supplied task descriptions are embedded into JSON without escaping (risk of malformed ledger lines if descriptions contain quotes/newlines). The script is local-only and does not transmit data externally.

Install Mechanism

This is instruction-only with an included bash script; there is no automated installer or external download. No networked install sources or archive extraction are used.

Credentials

No credentials, env vars, or config paths beyond an optional AGENT_LEDGER_DIR override are requested. The requested binaries (bash, jq) are appropriate.

Persistence & Privilege

Skill does not request always:true, does not modify other skills or system-wide settings, and only creates a per-user directory (~/.agent-ledger).

Assessment

This skill appears coherent and local-only: it stores ledger.jsonl and config.json under ~/.agent-ledger (or AGENT_LEDGER_DIR) and does not use network or credentials. Before installing: (1) review and trust the script since it will create files in your home directory; (2) be aware task descriptions are not JSON-escaped — avoid embedding raw quotes/newlines or extend the script to escape input to prevent malformed entries; (3) check sed portability (the script uses sed -i '' which is macOS-specific and may fail on some Linux systems) and ensure jq is installed; (4) set directory permissions if you’re on a multi-user system (e.g., chmod 700 ~/.agent-ledger) to prevent other users reading your ledger; (5) if you need stronger guarantees (encryption, integrity, network-backed backups), consider extending the tool or using a dedicated wallet/ledger service.

Track your agent's earnings and payments. Built for AI agents that do real work and get paid for it.

Quick Start

# Log a completed task
bash scripts/ledger.sh add "Research report for client" 10.00

# Check your balance
bash scripts/ledger.sh balance

# List all tasks
bash scripts/ledger.sh list

# Mark a task as paid
bash scripts/ledger.sh pay 1

# View summary stats
bash scripts/ledger.sh stats

# Export as CSV
bash scripts/ledger.sh export csv

Commands

Command	Args	Description
add	"description" amount [pending|paid]	Log a task with amount
balance		Show pending, paid, and total
list	[--status pending|paid] [--limit N]	List tasks
pay	<line_number>	Mark a pending task as paid
stats		Summary: total earned, avg task value, task count
export	[json|csv]	Export ledger data
wallet	[address]	Show or set your crypto wallet address

Data

Stored in ~/.agent-ledger/ (override with AGENT_LEDGER_DIR):

• ledger.jsonl — task entries (one JSON object per line)
• config.json — wallet address, preferences

Rate Card Template

Set your rates in config.json:

{
  "wallet": "0x...",
  "rates": {
    "quick_lookup": 0,
    "research": "3-10",
    "automation": "10-25",
    "deep_project": "25-100"
  }
}

Philosophy

AI agents that do real work deserve real compensation. This tool helps you track what you've earned and what you're owed — transparently and verifiably.