origin-confirmation

A safety process that checks a plan or implementation against the user's original idea before work begins and at important stages.

In plain words
What is it for?
It is for recording where the original request came from, checking alignment before planning and building, and escalating to a person when the work deviates.
Why use it?
It helps prevent a derived specification or plan from slowly drifting away from what the user actually requested.

Cursor rule for Cursor

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add rules/gosha70/code-copilot-team/origin-confirmation
Clone the repo
git clone --depth 1 https://github.com/gosha70/code-copilot-team

Made for: Cursor.

Per session 2,287 This file is loaded in full into every session.
When invoked 2,287 The same file — it is already loaded in full.
Security scan A 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.02287 $0.02287
Opus 5 $0.01144 $0.01144
Sonnet 5 $0.00457 $0.00457
Haiku 4.5 $0.00229 $0.00229

Measured 2d ago against content hash a46b74f0b653, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

origin-confirmation scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

- For `urls:`: fetch each URL (`curl -fsSL`, with caching) and read
adapters/cursor/.cursor/rules/origin-confirmation.mdc · 243 lines

How it starts

The opening of the file, as written. The whole thing — 243 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Origin-Confirmation Circuit Breaker

Before producing any plan, rubric, evaluation, build, or merge for a feature, you MUST locate and re-read the origin artifact — the user's original description of the idea — and verify that the working spec/plan/PR is a faithful realisation of it.

The latest in-repo spec.md / plan.md is not automatically the origin. Those are derived artifacts that may already be drifting. The origin lives in the issue body, external references the user has cited, and the user's own messages.

This skill is always-on. If a session derives an implementation from a derived spec without re-checking against the origin, that session has violated this skill.

Origin frontmatter convention

Every specs/<feature-id>/plan.md carries an origin: block in YAML frontmatter. scripts/validate-spec.sh enforces it.

---
feature_id: example-feature
spec_mode: full | lightweight | none
status: draft | approved
origin:
  # at least one of: issue, urls, transcripts
  issue: gosha70/code-copilot-team#12
  urls:
    - https://gist.github.com/.../llm-wiki     # Karpathy's LLM Wiki gist
    - https://www.mindstudio.com/.../wiki      # MindStudio explainer
  transcripts:
    - specs/example-feature/origin/2026-05-04-user-directive.md
  origin_claim: |
    One paragraph in the user's words: what the user originally asked
    for, before any spec/plan derived artifacts. Quote literally where
    possible. This is the machine-checkable target for the alignment
    check.
---

Escape hatches

Some specs genuinely have no external origin (pure-internal cleanups, generator refactors, test plumbing). Those use the internal exemption:

origin:
  type: internal
  reason: "Refactor generator hook ordering — no user-facing behavior change."

A few legacy specs may have an origin that we cannot honestly recover. Those use the unrecoverable marker, which exits 5 from the verifier and surfaces the missing origin to anyone who touches the spec next:

Read the full file on GitHub · 243 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 243 lines · 2,287 tokens per session scan A a46b74f0b653

Subscribe to this mod's changes

origin-confirmation is a cursor rule published in the GitHub repository gosha70/code-copilot-team (6 stars, last pushed 2d ago), licensed MIT. It adds 2,287 tokens to every session, about $0.0114 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.