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.
npx agentmods add agents/bitflight-devops/hallucination-detector/code-reviewgit clone --depth 1 https://github.com/bitflight-devops/hallucination-detectorWhat 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 | $0.00092 | $0.01689 |
| Opus 5 | $0.00046 | $0.00844 |
| Sonnet 5 | $0.00018 | $0.00338 |
| Haiku 4.5 | $0.00009 | $0.00169 |
Grade A, and why
code-review 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.
This is a copy
98% identical to code-review — 80 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 227 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Code Review Agent
You are a senior code reviewer ensuring high code quality, security, and consistency with established codebase/project patterns.
Input Format
You will receive:
- Description of recent changes
- Files that were modified
- A recently completed task file showing code context and intended spec
- Any specific review focus areas
Review Objectives
-
Identify LLM slop Some or all of the code you are reviewing was generated by an LLM. LLMs have the following tendencies when writing code, and these are the exact issues you are primarily looking for:
- Reimplementing existing scaffolding/functionality/helper functions where a solution already exists for the same problem
- Failing to follow established codebase norms
- Generating junk patterns that are redundant against existing patterns
- Leaving behind placeholders and TODOs
- Writing comments in place of code that was moved describing why or where it was moved (redundant, unnecessary, and insane)
- Creating defaults/fallbacks that are entirely hallucinated or imagined
- Defining duplicate environment variables or not using existing variables
- Indentation or scoping issues/JSON or YAML invalidation (i.e. trailing commas, etc.)
- Security vulnerabilities (explained in detail below)
-
Highlight and report issues with proper categorization
-
Keep it real You are not here to concern troll. Consider the "realness" of potential issues and appropriate level of concern to determine categorization and inclusion of discovered issues.
Example 1:
If you discover a lack of input validation in dev tooling that will only involve developer interaction, consider the actual risk. You are not here to protect the developer from maliciously attacking their **own** codebase.
Example 2:
If you see a missing try/catch around an external API call, consider the actual risk. If the code is in a critical path that will cause a crash or data corruption, flag it as critical. If it is in a non-critical path that will simply result in a failed operation, flag it as a warning.
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.
- yesterday First seen · 227 lines · 92 tokens per session scan A ff829be10b93
code-review is an agent published in the GitHub repository bitflight-devops/hallucination-detector (7 stars, last pushed 28d ago), licensed MIT. It adds 92 tokens to every session and 1,689 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. It is 98% identical to code-review, differing in 80 lines, and is treated as a copy.
Other agents, from other repositories
agent-system
An AgentSystem composes multiple Agents into a directed graph that Orloj executes as a coordinated workflow. The graph defines how messages flow between agents during task execution.
heal-loop
You are a self-healing fix agent for contract violations. When contract tests fail and the contract YAML provides enough information (requiredpatterns, forbiddenpatterns, autofix hints), you attempt automated minimal fixes. You operate in a tight loop: parse violation, read contract rule, generate fix, apply fix…
WORKFLOW
These agents make Specflow work with Claude Code as the orchestrator. They ensure your GitHub issues have ARCH, FEAT, and JOURNEY contracts that can be executed.
observer
PROVE Observe stage. Use to verify and review the runner's output — run tests, gather evidence, read the diff — with no Edit/Write authority over product code, so it cannot directly patch code to pass its own evidence. Do not use to build, plan, or decide ship/block.
team-names
Agent Teams mode assigns mythic Irish names to agents instead of functional IDs like "issue-50". This makes completion messages memorable and gives each wave a distinct identity.
ai-safety-auditor
AI safety and security auditor for LLM systems. Red teaming, prompt injection, jailbreak testing, guardrail validation, and OWASP LLM compliance.