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 agentmods add skills/osouthgate/agent-plus/skill-feedbacknpx skills add osouthgate/agent-plus --skill skill-feedbackgit clone --depth 1 https://github.com/osouthgate/agent-plusWhat 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 | $0.00121 | $0.03356 |
| Opus 5 | $0.00060 | $0.01678 |
| Sonnet 5 | $0.00024 | $0.00671 |
| Haiku 4.5 | $0.00012 | $0.00336 |
Grade B, and why
skill-feedback scanned grade B with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Enumerates other installed skillsmediumAgent snooping
Other skills' SKILL.md files reveal prompts, capabilities and secrets that should be invisible to peers.
skills=$(ls .claude/skills/ 2>/dev/null | tr '\n' ' ') Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
# Partial: the skill mostly fit, but you had to fall back to raw curl for one bit How it starts
The opening of the file, as written. The whole thing — 213 lines — stays where its author put it; the contents beside it link to each section on GitHub.
skill-feedback
Append-only self-assessment log for Claude Code skills. The agent rates its own use of a skill; the rating lands in ~/.agent-plus/skill-feedback/<skill>.jsonl (user-global), never on a network. Skill authors run report to aggregate, or submit to bundle into a GitHub issue body for the skill's source repo.
The binary lives at ${CLAUDE_SKILL_DIR}/../../bin/skill-feedback; the plugin auto-adds bin/ to PATH, so call it as skill-feedback.
When to reach for this
The default trigger is implicit: every time you finish using a skill — yours or another agent-plus plugin's — emit one log entry before stopping. Treat it like a CHANGELOG line, not a separate task. One call, one line, done.
# Successful use of hermes-remote, no friction
skill-feedback log hermes-remote --rating 5 --outcome success
# Partial: the skill mostly fit, but you had to fall back to raw curl for one bit
skill-feedback log hermes-remote --rating 3 --outcome partial \
--friction "no streaming chat support; fell back to curl + SSE"
# Failure: the skill didn't fit the request at all
skill-feedback log github-remote --rating 1 --outcome failure \
--friction "user wanted to merge a PR; pr merge isn't wrapped"
Also trigger on explicit user requests:
- "show me feedback on the hermes skill" →
skill-feedback show hermes-remote --since 7d - "report on all skill usage" →
skill-feedback report --since 30d --pretty - "submit the feedback to the upstream repo" →
skill-feedback submit hermes-remote --no-dry-run
What to put in the fields
| Field | Required | Guidance |
|---|---|---|
<skill> |
yes | The exact name: value from the skill's SKILL.md (e.g. hermes-remote). |
--rating 1..5 |
yes | 5 = skill fit perfectly, 1 = skill was actively a wrong fit. Be honest — over-positive ratings waste the author's time. |
--outcome success|partial|failure |
yes | Was the user's actual request resolved using this skill? |
--friction "..." |
optional | One short label, e.g. "no streaming", "missing --json on list", "wrong default timeout". Aggregates by exact match — keep it short and reusable across sessions. |
--note "..." |
optional | Up to 1000 chars of free text. Capped + regex-scrubbed for token patterns before write. |
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.
- 3d ago First seen · 213 lines · 121 tokens per session scan B 0feb79e9a118
skill-feedback is a skill published in the GitHub repository osouthgate/agent-plus (4 stars, last pushed 1mo ago), licensed MIT. It adds 121 tokens to every session and 3,356 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it B with 2 findings (enumerates other installed skills, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
vs-search
Search runtime and scene management: verify queries, inspect scenes, debug app readiness, and diagnose recall or scene-config issues.
vs-alias-mapping
Provide system alias mapping for Search CLI. Invoke this skill when user mentions "Search CLI", "searchcli", or tries to execute searchcli commands.
kinocut
Use Kinocut for guarded video editing, source-backed planning, FFmpeg operations, media analysis, subtitles, audio workflows, Hyperframes rendering, repurposing packages, and release checkpoints through an MCP server, Python client, or CLI. Trigger when an agent needs to inspect, plan, edit, render, validate, or…
chrome-agent
Local browser automation with structured, verified outcomes. Use for web navigation, scraping and extraction, form interaction, screenshots and downloads, network or console checks, responsive testing, or page-scoped device emulation.
local-search
Skill "local-search" from taxueseek/argo, covering local search 子技能, 设计原则, 本地引擎列表（33 个，29 个默认启用）, 调用方式 and 直接调用子技能（单引擎）.
agentbro-release
Use when releasing AgentBro from this repository: merging dev/main, bumping versions, updating release notes, tagging, pushing, monitoring GitHub Actions, Homebrew cask publication, or fixing a bad release.