Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/sd0xdev/sd0x-harnessnpx agentmods add skills/sd0xdev/sd0x-harness/pr-commentWrote 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/sd0xdev/sd0x-harness/pr-comment)<a href="https://agentmods.dev/skills/sd0xdev/sd0x-harness/pr-comment"><img src="https://agentmods.dev/badge/skills/sd0xdev/sd0x-harness/pr-comment/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/sd0xdev/sd0x-harness/pr-comment"><img src="https://agentmods.dev/badge/skills/sd0xdev/sd0x-harness/pr-comment.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00067 | $0.01491 |
| Opus 5.5 | $0.00027 | $0.00596 |
| Sonnet 5.5 | $0.00013 | $0.00298 |
| Haiku 4.5 | $0.00007 | $0.00149 |
Grade A, and why
pr-comment 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 1mo 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 — 179 lines — stays where its author put it; the contents beside it link to each section on GitHub.
PR Comment
Trigger Keywords
pr comment, review comment, post comment, send review, pr feedback, inline comment
When NOT to Use
| Need | Use Instead |
|---|---|
| Read existing PR review comments | /load-pr-review |
| Create a code review (automated) | /codex-review-fast or /codex-review |
| Create a PR | /create-pr |
| PR status overview | /pr-summary |
Core Principle
Discuss locally → prepare friendly comments → dry-run preview → user confirms → atomic submit
Control plane (this SKILL.md) handles conversation, tone, orchestration.
Data plane (JS script) handles validation, API calls, safety.
Tone Guidelines
Claude prepares comments following these rules:
| Rule | Description | Example |
|---|---|---|
| Questions over commands | Use "would it make sense to" instead of "change this to" | "Would it make sense to extract this to a shared utility?" |
| Code, not person | Subject is the code, not "you" | "This logic seems to..." not "You wrote..." |
| Explain why | Give the reason, not just the what | "If one copy gets a bug fix but the other doesn't..." |
| Assume good intent | Confirm rather than accuse | "Just want to confirm this trade-off is intentional" |
| Praise first | Acknowledge before suggesting | "This async design is great. One thought though..." |
| No emoji | Unless user explicitly requests | -- |
| Follow PR language | Match the language used in the PR | -- |
Internal classification (Conventional Comments, not shown in output):
| Label | Purpose | Blocking |
|---|---|---|
suggestion |
Improvement proposal | Configurable |
question |
Clarification | No |
issue |
Points out a problem | Yes |
nitpick |
Stylistic preference | No |
praise |
Positive feedback | No |
Workflow
sequenceDiagram
participant U as User
participant C as Claude (Conversation)
participant SK as SKILL.md (Orchestration)
participant JS as pr-comment.js (Data Plane)
participant GH as GitHub API
Note over U,C: Phase 1: Natural conversation
U->>C: Discuss PR diff, observations
C->>U: Draft N friendly comments
U->>C: Adjust tone / content / count
Note over U,GH: Phase 2: Skill invoked
U->>SK: /pr-comment [PR#]
SK->>SK: Collect comments from conversation
SK->>JS: prepare --pr N --repo owner/repo --input comments.json
JS->>GH: gh pr view (metadata)
JS->>GH: gh api pulls/N/files (changed files)
JS->>GH: gh api pulls/N (head SHA)
JS-->>SK: Validated payload + validation summary
SK->>U: Dry-run preview (default)
alt User confirms
SK->>U: AskUserQuestion gate
U->>SK: Approved
SK->>JS: submit --pr N --repo owner/repo --input payload.json
JS->>JS: Re-fetch head SHA, drift check
JS->>GH: POST /repos/{owner}/{repo}/pulls/{N}/reviews
GH-->>JS: 200/201 Success
JS-->>SK: Success + review URL
SK->>U: Review posted successfully
end
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 1mo ago First seen · 179 lines · 67 tokens per session scan A f06dac0f6259
pr-comment is a skill published in the GitHub repository sd0xdev/sd0x-harness (194 stars, last pushed 15d ago), licensed MIT. It adds 67 tokens to every session and 1,491 once invoked, about $0.0003 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 2026-09-03.
Other skills, from other repositories
comet-github-pr-review
A read-only review process for pull requests in Comet's GitHub repositories. A pull request is a proposed code change; the review checks its current code, related issue, comments, merge status, and automated checks.
update
Safe cc10x upgrade through the Claude Code plugin CLI. Lists every installed scope, refreshes the marketplace, updates each scope, and optionally carries local patches over to the new install. Use this skill when: updating cc10x, upgrading, pulling latest cc10x, syncing plugin, refreshing cache, or checking for new…
resolving-merge-conflicts
Use when a git merge or rebase reports conflicts and the operation is in progress.
code-review
Two-mode skill: (1) adversarial review — spec compliance + code quality + security, confidence-scored findings with file:line evidence; (2) receiving review — verify-before- agreeing discipline for acting on external/human review feedback.
codebase-hygiene
Two-mode skill: (1) find semantic duplicates — functions doing the same thing under different names, invisible to copy-paste detectors; (2) deepen shallow modules — thin wrappers and pass-through layers that spread complexity. Advisory and read-only; changes route through BUILD with full gates.
code-review
Review the changes since a fixed point (commit, branch, tag, or merge-base) along two axes — Standards (does the code follow this repo's documented coding standards?) and Spec (does the code match what the originating issue/spec asked for?). Runs both reviews in parallel sub-agents and reports them side by side. Use…