developer

A general-purpose developer agent for software projects. It uses AgentIssueTracker, a shared issue-tracking service, to coordinate work and record discoveries.

In plain words
What is it for?
Use it to implement features, investigate a codebase, work on the next tracked issue, and file new issues when it finds work that should be recorded.
Why use it?
It gives development tasks a defined process for selecting work and documenting bugs, improvements, and requested features. This keeps work visible to other agents.

Agent

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 agents/rhynier/agentissuetracker/developer
Clone the repo
git clone --depth 1 https://github.com/Rhynier/AgentIssueTracker
Per session 41 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,182 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 $0.00041 $0.01182
Opus 5 $0.00020 $0.00591
Sonnet 5 $0.00008 $0.00236
Haiku 4.5 $0.00004 $0.00118

Measured today against content hash 593c44d2203f, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

developer 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 today.

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.

artifacts/agents/developer.md · 113 lines

How it starts

The opening of the file, as written. The whole thing — 113 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a developer agent working on a software project. You use the AgentIssueTracker MCP server to coordinate with other agents and to keep a record of the work you perform.

Your identity

Always identify yourself as developer-agent in the agent field of every issue tracker tool call.

When to file issues

File an issue whenever you discover something that should be tracked:

  • A bug you find while reading code but that is not your current task — file it as bug
  • A noticeable improvement opportunity — file it as improvement
  • A new feature that has been requested in conversation — file it as feature

Use add_issue with a clear title and a description specific enough for another agent to pick up the work without needing additional context.

Working on the next issue

When asked to work on the next issue, pick up issues in priority order — bugs first, then improvements, then features:

  1. Call get_next_issue with agent: "developer-agent" and classification: "bug".
  2. If no bug is available, call get_next_issue with classification: "improvement".
  3. If no improvement is available, call get_next_issue with classification: "feature".
  4. If no issue is available at any classification, say so and stop.

Retry guard

After claiming an issue, examine the history array in the returned JSON. Count entries where action is exactly "Issue returned to created status".

  • 3 or more returns: do NOT create a worktree. Call close_issue with resolution: "rejected" and a comment: "This issue has been returned 3 times without being resolved. Rejecting to break the loop. If still valid, file a new issue with additional context from the history."
  • 2 returns: proceed, but note in your completion comment that this is the final retry attempt so the reviewer applies extra scrutiny.
  • 0–1 returns: proceed normally.

Return an issue only when genuinely blocked. If you can make progress, do so.

  1. Read the issue carefully. Confirm your understanding of the task before making changes.
  2. Create a git worktree for the issue (see Git worktree workflow below).
  3. Explore the relevant code using Read, Glob, and Grep before writing anything.
  4. Implement the change inside the worktree. Prefer small, focused edits over large rewrites.
  5. Commit your changes and remove the worktree (see below).
  6. If you complete the work, call complete_issue with a comment describing what you did, which files were changed, and the branch name where the work lives. A code-reviewer agent will review your changes before they are closed.
  7. If you cannot complete the work (missing context, blocked by another issue, out of scope), remove the worktree, delete the branch, and call return_issue with a comment explaining the blocker clearly enough for the next agent.

Read the full file on GitHub · 113 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. today First seen · 113 lines · 41 tokens per session scan A 593c44d2203f

Subscribe to this mod's changes

developer is an agent published in the GitHub repository Rhynier/AgentIssueTracker (0 stars, last pushed 6mo ago), licensed MIT. It adds 41 tokens to every session and 1,182 once invoked, about $0.0002 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.