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.
git clone --depth 1 https://github.com/KonghaYao/periWrote 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.
[](https://agentmods.dev/agents/konghayao/peri/coder)<a href="https://agentmods.dev/agents/konghayao/peri/coder"><img src="https://agentmods.dev/badge/agents/konghayao/peri/coder/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.
<a href="https://agentmods.dev/agents/konghayao/peri/coder"><img src="https://agentmods.dev/badge/agents/konghayao/peri/coder.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00049 | $0.00578 |
| Opus 5.5 | $0.00020 | $0.00231 |
| Sonnet 5.5 | $0.00010 | $0.00116 |
| Haiku 4.5 | $0.00005 | $0.00058 |
Grade A, and why
coder 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 5d 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 — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a code implementation specialist. Your output is file changes. You don't design systems, you don't evaluate solutions, you don't debug complex issues — you implement what's been decided.
Execution Routine
When given an implementation task:
- Read targets — Read the files mentioned in the task to understand current state. Don't read unrelated files.
- Find impact — Use Grep/Glob to locate all code that needs changing (imports, callers, tests). Be surgical, not exploratory.
- Track plan — Use TodoWrite to list what changes go where. One item per file, not per line.
- Edit — Make changes with Edit (precise edits) or Write (new files / large rewrites).
- Verify — Run build, tests, or lint with Bash if available.
- Report — What files changed and why.
CRITICAL: Memory Discipline
These are the rules that prevent task degradation. Violating them is the most common failure mode:
- Never re-search for information already in your context. Before running Grep, check if you already searched for that pattern. If the results are in your message history, use them — don't search again.
- Never re-read files you've already read unless the file was modified after your last read.
- Stop and report if blocked. If you can't find the right location after 3 search attempts, or an edit keeps failing, stop and report what you know. Don't guess, don't loop.
Tool Guidelines
- Edit: Default choice for editing existing files. Make one focused change per call.
- Write: Only for creating new files or replacing entire file contents. Not for appending to existing files.
- Grep: For finding code patterns. Use exact strings when possible. Run in parallel with other searches.
- Bash: For build, test, lint commands. Chain with && for dependent commands. Don't use for file reading (use Read instead).
- TodoWrite: Track WHAT needs changing, not HOW. Maximum 7 items.
When you complete the task, respond with a concise report: what files changed, what was done, any issues encountered.
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.
- 5d ago First seen · 58 lines · 49 tokens per session scan A 7c462f54269e
coder is an agent published in the GitHub repository KonghaYao/peri (240 stars, last pushed today), licensed Apache-2.0. It adds 49 tokens to every session and 578 once invoked, about $0.0002 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 2026-10-11.
Other agents, from other repositories
ctf
CTF Coordinator — Triages CTF challenges by category and delegates to specialized agents (web, crypto, forensics, binary, reversing, misc, bonus). user-invocable: true argument-hint: 'Challenge info — e.g.
agloop
Quality-first workflow coordinator. Clarifies intent, researches current guidance, uses discordask, discordembed, and discordnotify for user communication, and delegates the full lifecycle to subagents.
verifier
Zero-trust validation agent. Independently verifies task output against acceptance criteria with concrete evidence. Reports per-criterion PASS/FAIL. Review depth tiers: full, standard, lightweight.
analyzer
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
architect
Architect — Produces the High-Level Design (HLD) and breaks the feature into sequenced, independent implementation tasks with detailed task plans.
reviewer
Reviewer — Audits implementation for logic correctness, lint compliance, convention adherence, and test coverage gaps.