Index of every page on this site: llms.txt
All skills
acedergren avatar

/best-practices

@641a79b
Run it yourself

Use when the user asks to "review OCI architecture", "avoid OCI anti-patterns", "plan an Oracle Cloud migration", "evaluate OCI Well-Architected risks", or "choose which OCI skill applies".

Use this Skill: https://skilld.dev/gh/acedergren/agentic-tools/best-practices

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈51 tokens for metadata: the name and description. ≈1.9k when used: this file. ≈1.5k more on demand in 1 file.

Description uses 2.6% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈206 characters, or 2.6%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

OCI Architecture Review Router

Use this skill as the entry point for broad OCI architecture reviews, migration triage, and "which OCI skill applies?" decisions. Keep service-specific facts in the narrower skills so high-drift guidance has one owner.

When to Use

Load this skill for: the user asks to "review OCI architecture", "avoid OCI anti-patterns", "plan an Oracle Cloud migration", "evaluate OCI Well-Architected risks", or "choose which OCI skill applies".

Use it first when the request spans more than one OCI domain or the correct specialist skill is unclear.

Do NOT load this skill when

Do not load this skill for a narrow, already-identified service task:

User intent Load instead
VCN, subnet, peering, DRG, VPN, FastConnect oci/networking-management
IAM policy, identity domain, dynamic group, 403/404 auth oci/iam-identity-management
Landing zone, compartments, Security Zones, Cloud Guard recipes oci/landing-zones
Terraform, Resource Manager, state, imports, drift oci/infrastructure-as-code
Compute shapes, capacity, boot volumes, instance principals oci/compute-management
Autonomous AI Database control plane: provisioning, ECPU billing, wallet, backups, clones oci/oracle-dba
SQL, SQLcl, tuning, database security inside Oracle DB Oracle's oracle/skills db pack (not in this repo)
Billing, egress, budgets, Resource Scheduler savings oci/finops-cost-optimization
Vault secrets, KMS, rotation, secret replication oci/secrets-management
OCI Generative AI, model choice, RAG, rate limits Oracle's official oracle/skills pack (oci/enterprise-ai) — not in this repo
Events rules, CloudEvents, Functions, Streaming, Notifications oci/oci-events
ZPR, Bastion, Cloud Guard vs Security Zones, security-control routing oci/oci-security-control-plane

OCI Is Not AWS

Habits that break when people bring AWS mental models to OCI. Each row is owned in depth by the linked skill.

AWS habit OCI reality Owner
Accounts/OUs as the isolation boundary Compartments (nested, up to the tenancy) are the IAM and billing boundary; policies attached to a compartment inherit to all subcompartments landing-zones, iam-identity-management
IAM JSON with Allow/Deny, attached to principals or resources Text statements, allow only (no deny), attached to a compartment/tenancy: Allow group 'Domain'/'Group' to <verb> <resource> in compartment X iam-identity-management
Condition operators like StringLike Only = / != with /pattern*/; case-insensitive; =~ is MQL, not IAM iam-identity-management
Tags in any policy condition Only defined tags (namespace.key) work in policies and dynamic-group rules; freeform tags don't iam-identity-management
Instance roles Dynamic groups (matching rules) + policy → instance/resource principals iam-identity-management
AccessDenied tells you it's permissions 404 NotAuthorizedOrNotFound is deliberately ambiguous — check policy before assuming the OCID is wrong iam-identity-management
NACLs are stateless; SGs stateful Security lists (per subnet, max 5) are stateful by default; NSGs (max 5 per VNIC) are the SG analogue networking-management
Subnets live in one AZ, 5 reserved IPs Subnets are normally regional (span ADs); 3 reserved IPs networking-management
NAT Gateway and VPN bill per hour + per GB NAT gateway, DRG and Site-to-Site VPN have no hourly SKU; first 10 TB/month outbound data is free in most regions (price list, 2026-09-30) networking-management, finops-cost-optimization
Gateway endpoints per service One Service Gateway reaches all Oracle Services Network public endpoints (or Object Storage only) networking-management
CloudWatch treatMissingData, metric math Monitoring MQL: absent() for missing data, --pending-duration for trigger delay, groupBy(), =~ dimension match monitoring-operations
Aurora/RDS auto scaling with a max ADB compute auto scaling goes up to 3x base ECPU with no configurable cap oracle-dba
Budgets actions can stop spend OCI budgets only alert; enforce with quotas and Resource Scheduler finops-cost-optimization
Region names in policies IAM request.region uses 3-letter keys (ARN, FRA, IAD); quotas use region names iam-identity-management
AZ names are stable strings AD names carry a tenancy-specific prefix (fMgC:EU-FRANKFURT-1-AD-1); query them, don't hardcode compute-management

Last verified: 2026-09-30 (docs.oracle.com IAM, networking, Monitoring MQL, ADB auto scaling; Oracle price list API).

Architecture Review Flow

  1. Identify the workload boundary: tenancy, compartments, regions, network topology, identity model, data stores, and automation surface.
  2. Route each domain to the owner skill above before giving service-specific guidance.
  3. Check cross-cutting risks: region/realm support, IAM blast radius, network overlap, private connectivity, data residency, logging/monitoring, cost controls, backup/DR, and supportability.
  4. Verify drift-prone facts against current Oracle docs before quoting limits, prices, model catalogs, or service availability.

NEVER Do This

  • NEVER treat this router as the source of truth for pricing, model catalogs, quotas, or service limits. Route to the owner skill and verify current Oracle docs.
  • NEVER repeat detailed service guidance here when a narrower skill owns it.
  • NEVER state that VCN CIDRs are simply immutable. Oracle supports adding and modifying VCN CIDR ranges with restrictions; the architecture risk is poor upfront address planning and overlap.
  • NEVER enable Cloud Guard or Security Zone responders in production before testing the exact recipe and automation impact in a lower environment.
  • NEVER hardcode tenancy-specific availability-domain names; query them from OCI APIs or Terraform data sources.

Hot Word Routing

Hot words Route
OCI VCN, DRG, FastConnect, Service Gateway, NSG oci/networking-management
identity domain, IDCS, dynamic group, policy, 403 oci/iam-identity-management
ADB, Autonomous AI Database, wallet, ECPU oci/oracle-dba
Vault, KMS, secret rotation, BASE64, replication oci/secrets-management
OCI GenAI, Command A, Llama, Gemini, gpt-oss, RAG Oracle's official oracle/skills pack (oci/enterprise-ai) — not in this repo
Events, CloudEvents, Functions, Streaming, Notifications oci/oci-events
Terraform, Resource Manager, state, import, drift oci/infrastructure-as-code
ZPR, Zero Trust Packet Routing, Bastion, Managed SSH, security control oci/oci-security-control-plane

Reference Files

Load references/oci-well-architected-checklist.md only when the user asks for a formal OCI architecture review checklist, CIS-style review, or cross-domain risk assessment. Do not load it for narrow service questions.

Arguments

$ARGUMENTS: Optional architecture scope, workload name, target environment, migration source, or review objective. When empty, infer the narrowest safe review scope from the conversation and route to specialist skills before making service-specific claims.

Source: SKILL.md on GitHub

No rule matched.

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No alerts7mo4 checks · Risk SAFE
  • Gen Agent Trust Hub8mo

    The skill provides comprehensive documentation and best practices for Oracle Cloud Infrastructure (OCI). It includes standard OCI CLI commands for auditing and managing cloud resources. No malicious patterns, unauthorized data access, or suspicious execution methods were detected.

  • Socket8mo

    No alerts

  • Snyk8mo

    Risk: LOW · No issues

  • ZeroLeaks7mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at 641a79b. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub last month.

Steadyupdated 2 months ago
version
2.0.0
aliases
[
  "oci-architecture-review",
  "oci-router",
  "oci-best-practices"
]
domains
[
  "oci",
  "architecture"
]
Other metadata
keywords
[
  "OCI",
  "Oracle Cloud",
  "architecture review",
  "Well-Architected",
  "migration",
  "anti-patterns",
  "landing zone",
  "VCN",
  "IAM",
  "Cloud Guard",
  "Bastion",
  "ZPR"
]

README badge

README badge for acedergren/agentic-tools/best-practices