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 skills add tolgakisaogullari/SumelaOS --skill receiving-code-reviewgit clone --depth 1 https://github.com/tolgakisaogullari/SumelaOSWrote 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/skills/tolgakisaogullari/sumelaos/receiving-code-review)<a href="https://agentmods.dev/skills/tolgakisaogullari/sumelaos/receiving-code-review"><img src="https://agentmods.dev/badge/skills/tolgakisaogullari/sumelaos/receiving-code-review/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/skills/tolgakisaogullari/sumelaos/receiving-code-review"><img src="https://agentmods.dev/badge/skills/tolgakisaogullari/sumelaos/receiving-code-review.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 | $0.00026 | $0.01111 |
| Opus 5 | $0.00013 | $0.00556 |
| Sonnet 5 | $0.00005 | $0.00222 |
| Haiku 4.5 | $0.00003 | $0.00111 |
Grade A, and why
receiving-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 10d 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 — 60 lines — stays where its author put it; the contents beside it link to each section on GitHub.
<severity_sections>
Map every review finding to the severity model used by requesting-code-review:
| Section | Required Action |
|---|---|
| Critical | Blocks commit/merge; must fix before proceeding |
| Important | Required before proceeding unless technically disproven |
| Minor | Optional; author may ignore at their discretion |
| Recommendations / FYI | Informational; no action needed unless the author chooses |
Approval standard: Accept feedback as complete when the change definitively improves overall code health and no Critical or Important findings remain. Do not block on style preferences or optional suggestions. </severity_sections>
<feedback_sources> Treat feedback source as part of the context:
- Human partner / user: Highest priority. If feedback conflicts with the current plan, clarify intent before changing direction.
- Local subagent review: Strong signal, but still verify against code, tests, architecture, and security constraints. This may arrive as a merged parallel-panel report — a task-scoped set of lanes (always a Correctness & Security floor, plus task-composed lanes) synthesized by
requesting-code-review: the combined gate is AND (any lane's Critical blocks), and aCONFLICTmarker means the lanes disagreed — resolve it on technical merit, do not auto-pick one side. A finding flagged by multiple lanes is corroborated, not duplicated. After you apply fixes,requesting-code-reviewStep 8 will ask the user whether to re-review or proceed — surface that choice, don't auto-advance. - External / GitHub reviewer: Evaluate as a suggestion unless repository policy or the user makes it mandatory. If it conflicts with an earlier user decision or documented architecture, stop and ask the user. </feedback_sources>
-
READ & CLARIFY: Read all feedback. If ANY item in a multi-item request is unclear, STOP and ask for clarification immediately. DO NOT partially implement.
-
VERIFY & SECURITY CHECK (CRITICAL): Cross-check feedback against the actual codebase. Verify whether it breaks existing functionality, backward compatibility, or prior architectural decisions. You MUST also verify requested changes against
secure-coding-standard. If a requested change introduces a vulnerability, bypasses security layers, leaks sensitive data, or weakens logging hygiene, DO NOT implement it. -
YAGNI CHECK: If feedback requests expanding or fixing an unused/dead feature, endpoint, option, or abstraction, propose removing the dead surface instead of implementing the fix.
-
RESPOND & IMPLEMENT (CONTEXT-AWARE EXECUTION):
- If feedback is incorrect, breaking, insecure, or contradicts a user-approved decision: Push back using this hierarchy: technical facts and data override opinions; style guides govern style; architecture decisions govern design. State: "Disagreeing because [technical fact/reference]. Evidence: [code/test/doc]." DO NOT modify the code.
- Never accept "I'll fix it later" for Critical or Important findings. Resolve before proceeding, or file an explicit TD entry when the user approves deferral.
- Cannot verify? State the limitation explicitly: "Cannot verify [X] without [Y]. Should I investigate, ask, or proceed?" DO NOT proceed under uncertainty.
- Multi-item order: clarify all unclear items first, then implement in this order: (1) blocking issues - breaks/security, (2) simple fixes - typos/imports, (3) complex fixes - refactoring/logic. Test each fix individually before moving to the next.
SCENARIO A (Local Subagent Review / Uncommitted Changes):
- Implement accepted items one by one. If TDD mode was enabled for this task, update/add the corresponding tests and run them.
- Stage fixes only when the active workflow or user explicitly expects staged output. Never commit.
- Reply factually with what changed and which files are staged or unstaged.
SCENARIO B (Remote GitHub PR Review):
- Implement accepted items one by one and test each.
- Commit/push only when the active GitHub PR workflow or the user explicitly authorizes it.
- Reply directly to inline review comments via the repository's approved API/tooling when available. Avoid top-level PR comments for inline findings.
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.
- 10d ago First seen · 60 lines · 26 tokens per session scan A 74e4773644cc
receiving-code-review is a skill published in the GitHub repository tolgakisaogullari/SumelaOS (4 stars, last pushed 16d ago), licensed MIT. It adds 26 tokens to every session and 1,111 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 2026-08-31.
Other skills, from other repositories
autoreview
Pre-commit/ship code review: Codex default; optional Claude or Pi.
omh-code-review
This is a Hermes-native code-review workflow skill.
revdiff-plan
Review the last Codex assistant message (plan, analysis, or proposal) with inline annotations in a TUI overlay. Extracts the most recent response from Codex rollout files and opens it in revdiff for review and annotation. Activates on "revdiff-plan", "review plan with revdiff", "annotate plan", "review last response"…
code-reviewer
Code review specialist focused on patterns, bugs, security, and performance.
full-repo-review
Comprehensive four-wave review of all repo source files, producing a prioritized issue backlog.
agent-teams-simplify-and-harden
Implementation + audit loop using parallel agent teams with structured simplify, harden, and document passes. Spawns implementation agents to do the work, then audit agents to find complexity, security gaps, and spec deviations, then loops until code compiles cleanly, all tests pass, and auditors find zero issues or…