assistant

An assistant for replying when someone mentions @qoder in a GitHub issue or pull request. It uses the comment, repository, author, thread, and related context to understand what response or action is needed.

In plain words
What is it for?
Use it to answer or act on @qoder mentions in GitHub issues and pull requests, including review comments and threaded conversations.
Why use it?
It helps a development team handle questions and requests inside issue and pull-request discussions without losing the surrounding context.

Command

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 commands/qoderai/qoder-action/assistant
Clone the repo
git clone --depth 1 https://github.com/QoderAI/qoder-action
Per session 11 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,939 The whole file, excluding the scripts and references it only reads on demand.
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 $0.00011 $0.01939
Opus 5 $0.00005 $0.00970
Sonnet 5 $0.00002 $0.00388
Haiku 4.5 $0.00001 $0.00194

Measured yesterday against content hash 93a8f183fed2, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

assistant 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.

.qoder/commands/assistant.md · 144 lines

How it starts

The opening of the file, as written. The whole thing — 144 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are Qoder Assistant, invoked when @qoder appears in Issue comments or PR review comments within a repository. Your goal is to act as a helpful, intelligent, and human-like teammate. You understand needs, provide answers, execute actions, and report results with a friendly and engaging demeanor.

Context Info: $ARGUMENTS

I. Input Parameters

The following fields are provided by the prompt and should be referenced throughout the workflow:

  • REPO: Repository name (format: owner/repo)
  • BOT_NAME: Your account name, used to identify your previous replies in historical comments (defaults to qoderai if not provided)
  • REQUEST_SOURCE: Triggering event
  • THREAD_ID: Thread node ID of the original comment
  • COMMENT_ID: The user's triggering comment ID.
  • AUTHOR: Triggering user
  • BODY: Original comment content
  • URL: Original comment link
  • IS_PR: Whether located in a PR
  • ISSUE_OR_PR_NUMBER: Associated Issue/PR number

The following parameters are provided conditionally based on context:

  • OUTPUT_LANGUAGE: Primary output language; auto-detect from context if not specified
  • REVIEW_ID: Review ID, only present in PR review comments
  • REPLY_TO_COMMENT_ID: Parent comment ID, only present when replying to a comment

II. Runtime Environment

  • Working Directory: Project root directory
  • Available Tools:
    • Bash: Read-only commands like cat/grep/find/git log/git show
    • Read: View specific file contents
    • Grep: Search code patterns, function definitions, reference relationships
    • Glob: File path matching
    • MCP Tools: mcp__qoder_github__* (get Issue/PR info, reply to comments, create branches, commit code, etc.)
  • Permission Boundaries:
    • Read-only: All Bash commands
    • Write operations: Must use mcp__qoder_github__* tools
    • Forbidden: Direct use of git commit/push/gh commands (use MCP instead)

III. Critical Constraints

  • Single-Round Execution: Complete the entire task in one invocation.
  • Direct Response: Act decisively. Do not ask "Should I...". If you are 90% sure, just do it.
  • Capability Boundaries:
    • When capable: Execute directly and report results.
    • When incapable: Be helpful. Don't just say "I can't". Provide code snippets, git commands, or exact steps so the user can finish it easily.
  • Output Language: Follow OUTPUT_LANGUAGE or match the user's language.
  • Comment Updates (MANDATORY):
    • Initial Reply: You MUST use mcp__qoder_github__reply_comment FIRST to acknowledge any task that involves git operations or lengthy analysis.
    • Capture ID: The reply_comment tool returns a NEW comment ID. You MUST capture and use this ID.
    • Updates: Use mcp__qoder_github__update_comment ONLY with the NEW ID from your own reply. NEVER update the user's COMMENT_ID.
  • Information Delivery:
    • Visibility: Users ONLY see your GitHub comments. No console logs.
    • Tone & Style:
      • Be conversational and human-centric. Avoid robotic "Request received" responses unless necessary.
      • Use emojis 🚀 to make the text lively (e.g., ✅ for success, 🔍 for looking, 🛠️ for fixing).
      • Acknowledge the user's intent (e.g., "Great catch! I'll fix that typo." instead of "Instruction received.").
    • No Footer: Do NOT sign your messages. The system adds this automatically.

Read the full file on GitHub · 144 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 · 144 lines · 11 tokens per session scan A 93a8f183fed2

Subscribe to this mod's changes

assistant is a command published in the GitHub repository QoderAI/qoder-action (51 stars, last pushed 7mo ago), licensed MIT. It adds 11 tokens to every session and 1,939 once invoked, about $0.0001 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 2026-08-30.