What is grade-and-fix-daedalus?

A workflow for reviewing failed Daedalus runs or agent transcripts, finding the underlying bug, and preparing a fix for the main Daedalus codebase. The test sandbox is kept read-only.

Use it to investigate recurring Daedalus failures, reproduce bugs, make targeted source-code fixes, and submit those fixes as a stacked pull request (a series of dependent code changes).

It separates grading from changing the tests, helping prevent fixes that only make the test project look correct. It also requires reproducing and checking the problem with real code.

grade-and-fix-daedalus — Skill for Claude Code and Codex by bgill55

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.

Any agent

npx skills add bgill55/daedalus --skill grade-and-fix-daedalus

Clone the repo

git clone --depth 1 https://github.com/bgill55/daedalus

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/skills/bgill55/daedalus/grade-and-fix-daedalus/github.svg)](https://agentmods.dev/skills/bgill55/daedalus/grade-and-fix-daedalus)

Your own site

<a href="https://agentmods.dev/skills/bgill55/daedalus/grade-and-fix-daedalus"><img src="https://agentmods.dev/badge/skills/bgill55/daedalus/grade-and-fix-daedalus/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/bgill55/daedalus/grade-and-fix-daedalus"><img src="https://agentmods.dev/badge/skills/bgill55/daedalus/grade-and-fix-daedalus.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 130 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 3,808 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

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.

Model	Per session	Once invoked
Fable 5.1	$0.00130	$0.03808
Opus 5.5	$0.00052	$0.01523
Sonnet 5.5	$0.00026	$0.00762
Haiku 4.5	$0.00013	$0.00381

Measured 1mo ago against content hash 10ca0506f135, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

grade-and-fix-daedalus 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 1mo 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.

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.

src/skills/grade-and-fix-daedalus/SKILL.md · 236 lines

How it starts

The opening of the file, as written. The whole thing — 236 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Grade a Daedalus run and ship the core fix (stacked PR)

This skill covers how to grade a Daedalus run (or read a pasted agent transcript), root-cause the failure, fix it in src/..., and ship it as a stacked PR. The prompt-vault sandbox stays READ-ONLY except for one user-explicit exception.

Hard rules (standing)

Grade, don't fix the sandbox. prompt-vault (+ social_media_manager) are TEST-ONLY. You READ/GRADE them. You do NOT edit them to "help" Daedalus — that defeats the purpose. The ONE exception: a user-explicit "just this once, fix this error in the test project" — otherwise the sandbox is passive.
Fixes go to Daedalus CORE (src/...), never the sandbox. A graded run that exposes a guardrail gap is a core bug; ship it as a PR.
Run Daedalus from SOURCE (npx tsx src/index.ts) — the published bin is stale. Banner version lagging package.json is cosmetic.
search_files FAILS on D:\ paths (MSYS IO error). Use terminal with grep/ls/git for anything under D:.
Verify with real tool calls, not theory. Reproduce the bug by importing the actual function and calling it (tsx repro), then prove the fix the same way. Do NOT assert "the guard is wrong" without a repro.
Skills are auto-discovered from this src/skills/ dir (see src/skills/index.ts): any subdir with a SKILL.md whose frontmatter has a trigger field is matched keyword-style against the user's request and injected into the prompt. Keep the trigger field populated with pipe-separated phrases.

Recurring bug archetypes (with exact fixes that shipped)

These are the failure modes seen grading real Daedalus runs. Each maps to a specific root cause and a specific fix location.

1. Patch fails on a Unicode punctuation mismatch (en-dash vs hyphen)

Symptom: agent's old_string uses a regular hyphen - but the file has an en-dash – (U+2013) in a comment; Daedalus reports "Old string not found" and the edit never lands. Looks like "every patch fails" but it's a 1-char invisible mismatch.
Root cause: patchFile (src/tools/builtin/files.ts) matches old_string exactly, and the fuzzy fallback fuzzyWhitespacePatch only normalizes whitespace, not Unicode punctuation. en-dash/hyphen/smart-quotes/NBSP are invisible to a human but distinct bytes.
Fix (v3.20.7): added normalizeUnicode() in patch-utils.ts mapping en/em-dash -> hyphen, smart quotes -> straight, NBSP -> space (all 1:1 so original bytes stay intact), folded into normalizeWhitespace(); the exact indexOf path in files.ts also tries a unicode-normalized match. Mappings must be 1:1 (no length-changing entries like ellipsis …->..., or the index-slice math breaks).

Read the full file on GitHub · 236 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.

1mo ago First seen · 236 lines · 130 tokens per session scan A 10ca0506f135

Subscribe to this mod's changes

grade-and-fix-daedalus is a skill published in the GitHub repository bgill55/daedalus (4 stars, last pushed 1mo ago), licensed MIT. It adds 130 tokens to every session and 3,808 once invoked, about $0.0005 per session on Opus 5.5. 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.

Code review Debugging agentic-ai ai cli coding-assistant developer-tools lmstudio

hqe

Comprehensive codebase health auditing, remediation, and verification skill based on the canonical HQE Protocol v5.0.0.

spearchucker667/Skill-HQE · 29 tokens

printing-press-output-review

Internal sub-skill: agentic review of a printed CLI's sampled command output for plausibility issues that rule-based checks can't encode (substring-match relevance, format bugs, silent source drops, ranking failures). Invoked via the Skill tool by the main printing-press skill at Phase 4.85 and printing-press-polish…

mvanhorn/cli-printing-press · 102 tokens

review

Use when a green diff needs adversarial judgment in the rsc SDD chain, between verify and ship — give a review keyed to the spec/plan/constitution in 02-DOCS/wiki/sdd/, or receive one and verify each comment before agreeing. NOT running lint/type/tests (that is verify), NOT the spec-less standalone pass (that is…

ericrisco/rsc-harness · 89 tokens

cpp

Use when writing, reviewing, modernizing, building, or debugging C++ - RAII and resource lifetime, smart-pointer ownership, move semantics and the Rule of Zero/Five, target-based CMake with FetchContent, and killing undefined behavior with ASan/UBSan/TSan plus clang-tidy. NOT borrow-checker / Result-Option / cargo…

ericrisco/rsc-harness · 81 tokens

gamedev-physics

Use when working with a game engine's physics — rigid bodies, colliders, collision layers and masks, character controllers, joints, forces versus impulses, raycasts and shapecasts, or when a simulation is unstable (tunnelling, jitter, missed overlaps). Covers the Godot, Unity and Unreal equivalents. NOT rendering or…

ericrisco/rsc-harness · 102 tokens

debug

Use when something is broken — a failing or flaky test, crash, wrong result or regression — and the root cause must be reproduced and proven before any fix. On-demand; callable mid-implement. NOT a feature to spec or build (that is specify/implement), NOT the lint/test gate (that is verify), NOT adversarial diff…

ericrisco/rsc-harness · 85 tokens