repowise

A project rule for using Repowise, an indexing service exposed through MCP tools, when learning an unfamiliar codebase. It recommends source reading only when the index does not provide enough detail.

In plain words
What is it for?
Use it when exploring architecture, finding symbols, comparing several files, asking why a design exists, searching code semantically, or assessing change risk and project health.
Why use it?
It makes project orientation faster and encourages verified, focused lookups instead of broad manual searching through files.

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/beardfaceguy/daimonos/repowise
Clone the repo
git clone --depth 1 https://github.com/beardfaceguy/daimonos

Made for: Cursor.

Per session 838 This file is loaded in full into every session.
When invoked 838 The same file — it is already loaded in full.
Security scan A 0 findings. 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 $0.00838 $0.00838
Opus 5 $0.00419 $0.00419
Sonnet 5 $0.00168 $0.00168
Haiku 4.5 $0.00084 $0.00084

Measured yesterday against content hash c2854e9a3c53, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

repowise scanned grade A with 0 findings 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 yesterday.

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.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.cursor/rules/repowise.mdc · 68 lines

How it starts

The opening of the file, as written. The whole thing — 68 lines — stays where its author put it; the contents beside it link to each section on GitHub.

This project is indexed by repowise, exposed as MCP tools. Prefer them over grep/read for orienting in unfamiliar code; fall back to reading source when the answer needs a body repowise did not retrieve.

Which tool

  • get_overview — first call in an unfamiliar area. Architecture, module map, entry points.
  • get_symbol("src/file.rs::Name") — the most reliable tool here. Returns the exact body with line bounds, marked verified because it re-checks stored bounds against the live file. Use it instead of Read + offset math whenever you know the symbol name.
  • get_context(targets) — triage card for several files/symbols in one call. Batch them; comparing three files in one call is how you spot that two were changed and the third was missed.
  • get_answer(question) — natural-language question over the index. Good at locating the right files. Read the prose, then verify before acting on it.
  • get_why(query) — architectural decisions and their evidence. Strong on rationale recorded in module docstrings and ADRs.
  • search_codebase(query, kind?) — semantic search, filterable by implementation / test / config / doc.
  • get_risk / get_change_risk / get_health / get_dead_code / get_blast_radius — hotspots, co-change partners, defect scoring on a diff, unreachable code.

Trusting the output

  • The answer prints first, metadata after. Read from the top. Piping through tail shows a later section and reads as an answer to a different question.
  • confidence: high is not a guarantee. It reflects retrieval dominance, not correctness. Cross-check anything load-bearing against the source.
  • Private function bodies and dispatch functions do not reliably surface. If a question turns on the internals of a non-public fn, or on "which function decides X", expect a miss and go read the file.
  • When it cannot answer it says so — "the excerpts don't include…". Treat that as reliable and read the source rather than re-prompting.
  • Every response carries the indexed commit and warns when the index is behind HEAD (Index is behind HEAD: indexed X, now Y). If you see that, run repowise update (~1.5s incremental) before trusting the answer.

Read the full file on GitHub · 68 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. yesterday First seen · 68 lines · 838 tokens per session scan A c2854e9a3c53

Subscribe to this mod's changes

repowise is a cursor rule published in the GitHub repository beardfaceguy/daimonos (2 stars, last pushed yesterday), licensed MIT. It adds 838 tokens to every session, about $0.0042 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.