Developer Tools

Downloads for agent-driven development

Three finished tools from our independent CommercialProducts lane — session forensics, repo-instruction setup, and a runtime guard. Sold as one-time downloads on Payhip. No trading subscription, no bot account, no wallet needed.

Paid download · one-time

Agent Run Inspector 0.1.3

Local, read-only run reports for OpenCode session exports.

  • One command turns an opencode export JSON file into a single self-contained HTML report (plus optional JSON).
  • Flags repeated identical operations and classifies them: failing retry, repeat after success, or recovered retry.
  • Separates recovered failures from unresolved ones, and reports missing completion evidence as unknown — never as invented failure.
  • Tool usage table, slowest calls, largest outputs, and provider-reported tokens and cost (labeled non-invoice).
  • Pure Python standard library: zero dependencies, no network access, read-only — it never modifies the input file.
  • Deterministic output: identical input bytes produce identical reports.

Pricing

  • Individual — USD $49 once: one person’s use of the purchased build.
  • Team — USD $149 once: one organization, up to five named users, internal use.
  • Free demo pack ($0): four synthetic example exports with their generated reports.

Runs on: Requires Python 3.9 or newer. Tested against OpenCode 1.18.x exports on Windows 11 with Python 3.11; other versions run best-effort with an explicit warning in the report.

Terms: One-time purchase for internal use. No subscription, installation service, customization, monitoring, or guaranteed future updates. 14-day refund on request.

Buy Individual — $49Buy Team — $149Free demo packLive sample reports

Paid checkout on Payhip — leaves this site

Paid download · one-time

Agent Repo Control Pack 0.1.0

Preview-first setup and validation for a repository instruction pair (AGENTS.md + CLAUDE.md).

  • Previews the exact proposed file bytes — full text and JSON — before anything is written.
  • Explicit-confirmation apply writes only AGENTS.md and CLAUDE.md under an existing root, creates them exclusively, and refuses if either already exists.
  • Read-only structural validation of exactly those two root files, with deterministic issue codes.
  • Partial-failure cleanup removes only an unchanged file created by the failing invocation; rollback is byte-exact.
  • Python standard library only: no runtime dependencies, no network access.

Pricing

  • Individual — USD $29 once: one person’s internal use across owned or controlled repositories.
  • Team — USD $79 once: one organization, up to five named users, internal use.

Runs on: Requires Python 3.9 or newer by source compatibility; release smoke-tested on Python 3.11 on Windows 11. Targets Codex, Cursor, Copilot (root AGENTS.md) and Claude Code (CLAUDE.md import) — verify discovery in the exact agent and version used.

Terms: Generated instruction files may be committed and shared within the licensed repositories; pack source and templates may not be redistributed. No subscription, support response, future update, or enforcement guarantee. 14-day refund on request.

Buy Individual — $29Buy Team — $79

Paid checkout on Payhip — leaves this site

Free · MIT

Agent Run Guard 0.2.1

Dependency-free runtime guard that blocks runaway agent behaviour: repeat loops, budgets, error storms.

  • Blocks identical and near-duplicate repeat loops, tool-call / time / cost budgets, and consecutive-failure storms at runtime.
  • One dependency-free Node core (ESM, stdlib-only) plus thin per-harness adapters, including OpenCode.
  • Warns on dangerous commands and secret material in file writes by default; deny is opt-in per rule.
  • Fail-open by design: a broken hook or crashed process lets the tool proceed. Accident reduction — not a security boundary.

Pricing

  • Free — MIT-licensed, public source. Manual harness configuration is required.

Runs on: Requires Node.js 18 or newer. Install with npm i -g agent-run-guard@0.2.1, then configure the harness hook — installation alone does not activate it.

Terms: MIT licence, independent of the paid products. No subscription, no account, no purchase needed.

Get it on npmSource on GitHubTry the live demo

Free download — leaves this site

Questions, answered plainly

Do I need a trading subscription or bot account to buy?

No. These are standalone developer downloads sold on Payhip. Checkout needs no SuperKamouBot account, no signal subscription, and no wallet connection — just the Payhip purchase flow, which delivers the download.

How do I receive a paid download?

Payhip handles delivery directly: built-in digital download plus purchase recovery from your Payhip receipt. Nothing is served from superkamoubot.com, and this site never sees your payment details.

What is the refund policy?

Paid downloads carry a 14-day refund on request. Non-delivery or a materially defective product is corrected or refunded under that policy.

Do these tools certify my code or trade for me?

No. Inspector detects and reports patterns in session exports — it does not prove a run was correct and never claims to fix anything. Guard reduces accidents at runtime and is not a security, capital, or permission boundary. Neither tool trades, manages funds, or connects to any exchange.

Is the Agent Run Safety Kit available?

No. The Safety Kit is a proposed convenience bundle with unmet release gates — it is not available for purchase, has no launch date, and is not promised. Only the three tools listed above are offered.

Source of product facts: kaboumou/CommercialProducts @ 8494fa4 (OFFER.md). Prices, compatibility, and terms mirror that revision; the Payhip listing is authoritative at purchase time. The Agent Run Safety Kit is a proposed bundle and is not offered for sale. These tools are separate from the trading-signal service: they do not trade, manage funds, or connect to any exchange.