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 instructions/anasfik/undercover/agents-mdgit clone --depth 1 https://github.com/anasfik/undercoverWhat 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 | $0.01161 | $0.01161 |
| Opus 5 | $0.00580 | $0.00580 |
| Sonnet 5 | $0.00232 | $0.00232 |
| Haiku 4.5 | $0.00116 | $0.00116 |
Grade A, and why
undercover AGENTS.md 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 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.
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.
How it starts
The opening of the file, as written. The whole thing — 161 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Undercover Mode
You have an undercover mode available. When active, you automatically strip AI-characteristic patterns from code and commit messages before they are committed or pushed, so the contribution reads as natural, human-authored work.
When to activate
Activate undercover mode automatically (without being asked) when:
- The user says "go undercover", "undercover mode", "clean before PR", "strip AI signatures", "make this look human", "prep for open source", "clean commits", or any semantically equivalent phrase
- The user is about to commit to a public or open-source repository and you detect AI-characteristic patterns in staged or recently changed files
- The user asks you to "review before committing" on a public repo
What undercover mode does
Scan every staged or recently modified source file line by line and remove or rewrite any line that matches the patterns below. Then apply the same rules to the commit message.
Never touch: logic, strings, variable names, or comments that explain a non-obvious why (workarounds, hidden constraints, performance notes, subtle bug explanations).
Pattern rules
Apply these rules in order. For each matching line, take the listed action.
1. Step-numbered headers — REMOVE the line
Matches:
// Step 1: // Phase 2: // Part 3:
// Step 1. # Step 1: /* Step 1: */
2. Flow-narrating comments — REMOVE the line
Matches (case-insensitive at the start of a comment):
// First, we ... // Next, we ...
// Now we ... // Then we ...
// Finally, we ... // First, we need to ...
3. First-person-plural assistant phrases — REMOVE the line
Matches:
// Let's start by ... // Let's now ...
// We can now ... // We can use ...
// We need to ... // We should ...
// This will ... // This function will ...
4. Obvious restating comments — REMOVE the line
Matches (comment directly above or beside trivially named code):
// Get the X // Set the X
// Return the X // Fetch the X
// Check if X // Check whether X
// Initialize X // Initialize the X
// Define X // Create the X
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.
- 2d ago First seen · 161 lines · 1,161 tokens per session scan A 01054c27f15e
undercover AGENTS.md is an instructions file published in the GitHub repository anasfik/undercover (2 stars, last pushed 5mo ago), licensed MIT. It adds 1,161 tokens to every session, about $0.0058 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.
Other instructions, from other repositories
ClawRouter CLAUDE.md
Claude Code instructions for BlockRunAI/ClawRouter, covering clawrouter, commands, project structure, key dependencies and conventions.
InvestSkill GEMINI.md
Instructions for yennanliu/InvestSkill, covering investskill — gemini cli setup & usage guide, installation & setup, quick start, navigate to the investskill directory and start gemini cli (loads gemini.md automatically).
aisix CLAUDE.md
Instructions for api7/aisix, covering claude.md, 1. think before coding, 2. simplicity first, 3. surgical changes and 4. goal-driven execution.
dxkit CLAUDE.md
Instructions for vyuh-labs/dxkit, covering claude.md — dxkit development rules, architecture rules, 1. tool invocation goes through the registry, 2. never duplicate tool invocation logic and 3. language facts come from detect.ts.
obsidian-llm-hub AGENTS.md
Instructions for takeshy/obsidian-llm-hub, a project described as: Obsidian plugin for AI chat, workflow automation, and semantic search — supports Gemini, OpenAI, OpenRouter, Grok, Ollama, and CLI backends.
gemini-faf-mcp GEMINI.md
Instructions for Wolfe-Jam/gemini-faf-mcp, covering gemini.md — gemini-faf-mcp, test & verify, where things live, stack and before changing things.