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 orcasecurity/orca-skills --skill orca-overprivileged-identities-rightsizinggit clone --depth 1 https://github.com/orcasecurity/orca-skillsWrote 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/orcasecurity/orca-skills/orca-overprivileged-identities-rightsizing)<a href="https://agentmods.dev/skills/orcasecurity/orca-skills/orca-overprivileged-identities-rightsizing"><img src="https://agentmods.dev/badge/skills/orcasecurity/orca-skills/orca-overprivileged-identities-rightsizing/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/orcasecurity/orca-skills/orca-overprivileged-identities-rightsizing"><img src="https://agentmods.dev/badge/skills/orcasecurity/orca-skills/orca-overprivileged-identities-rightsizing.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 2 findings, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 144 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
- medium Excessive Agency · line 218 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
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 | $0.00099 | $0.09342 |
| Opus 5 | $0.00049 | $0.04671 |
| Sonnet 5 | $0.00020 | $0.01868 |
| Haiku 4.5 | $0.00010 | $0.00934 |
Grade A, and why
orca-overprivileged-identities-rightsizing 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 11d 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 — 249 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Orca Over-Privileged Identity Right-Sizing Skill
Answers the question: "Which of our identities hold far more permission than they use, and how do we safely cut them down?"
Every unused permission is standing attack surface: if the identity is phished or its key leaks, the blast radius is everything it may do, not what it does do. This skill sweeps an account or business unit for identities whose granted permissions exceed their observed usage, ranks them by identity risk score, and walks the user through right-sizing with a non-destructive path (stage: generate the change, apply nothing) and a destructive path (apply) that is always gated behind an evidence-based safety check plus explicit confirmation.
The core signal: Orca's recommendation engine pre-computes a PoLP verdict per identity/grant from ~90 days of observed usage, stored as RecommendationType with exactly three values: "Reduce Permissions" (over-privileged — this skill's target), "Inactive" (dead weight — hand off to /orca-inactive-identities-cleanup), and "PoLP Aligned" (nothing to do). Where the verdict lives differs per provider:
- AWS (identity-level): an
AwsEffectivePermissionsPolicyper user/role carries the verdict plus a generated least-privilege policy (Recommendation.recommended_policy),UsedServicesvsEntityAuthorizedServices, and the identity'sPermissionUsageratio (share of authorized services actually used). Exposed directly viaget_aws_effective_permissions_policy_on_asset. The engine skips IAM groups. - Azure (grant-level): each
AzureIamRoleAssignmentcarries both the normalizedRecommendationTypeverdict and an inlineRecommendationpayload: a typed action (detach_role,read_only,just_in_time,scope_reduction,no_action_needed), anaction_neededboolean, and read/write action usage withLastUsageTime. The typed actions roll up to the three verdicts (detach_role→ "Inactive";read_only/just_in_time/scope_reduction→ "Reduce Permissions";no_action_needed→ "PoLP Aligned"). Computed only where cloud logs are enabled for the subscription. - GCP (grant-level): a
GcpIamPolicyBindingRecommendationper policy binding, same normalized-verdict-plus-typed-payload shape, linked to the user or service account. Feature-flag gated in some tenants.
Provider coverage is a product fact: Alibaba Cloud, OCI, and Tencent Cloud have no PoLP recommendation layer in Orca. This is a three-cloud skill (AWS, Azure, GCP); say so instead of improvising verdicts for other providers.
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.
- 11d ago First seen · 249 lines · 99 tokens per session scan A 636a3f7db985
orca-overprivileged-identities-rightsizing is a skill published in the GitHub repository orcasecurity/orca-skills (50 stars, last pushed 5d ago), licensed MIT. It adds 99 tokens to every session and 9,342 once invoked, about $0.0005 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 2026-08-30.
Other skills, from other repositories
happiness-skill
A Chinese-language guide to happiness based on reducing unmet wants, focusing on the present, and treating happiness as a trainable skill.
docx-comment-reply
Reply to comments (批注) in Word .docx/.doc files: extract comment context, draft replies, write threaded replies back, and validate OOXML.
asc-subscription-localization
Bulk-localize subscription, subscription-group, and in-app purchase display names across App Store locales using asc, including API 4.4.1 version-scoped v2 resources. Use when filling or updating subscription/IAP names and descriptions without App Store Connect UI work.
pcbway
PCBWay PCB fabrication and assembly — turnkey/consigned assembly, design rules, ordering workflow. Alternative to JLCPCB for manufacturing. Use with KiCad. Use this skill when the user mentions PCBWay, needs turnkey assembly (PCBWay sources parts by MPN), has parts not available on LCSC, needs assembled boards with…
explaining-machine-learning-models
Explain trained machine learning models through feature attribution, local explanations, and behavior summaries. Use as an explicit/manual helper once a model already exists, not for training ownership, leakage auditing, or general ML strategy selection.
influence-psychology
Apply the seven principles of ethical persuasion (reciprocity, commitment, social proof, authority, liking, scarcity, unity) to product design, copy, and sales. Use when the user mentions "social proof", "persuasive copy", "why users dont convert", "ethical persuasion", "reciprocity", "scarcity tactics", "commitment…