AgentStack
SKILL verified MIT Self-run

Oci Cost

skill-adibirzu-oci-skills-oci-cost · by adibirzu

>-

No reviews yet
0 installs
14 views
0.0% view→install

Install

$ agentstack add skill-adibirzu-oci-skills-oci-cost

✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

Are you the author of Oci Cost? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

OCI Cost & Usage (FinOps)

Answer "what is this tenancy costing, and where?" without learning the Usage API. This skill is read-first — it reports spend and budgets and recommends guardrails. It leans on the shared tenancy-safety core (oci_cli, redact).

First move (always)

  1. Confirm which tenancy you are reading (cost is tenancy-scoped):

``bash ./scripts/oci_preflight.sh ``

  1. Get the spend + budget snapshot in one call:

``bash ./scripts/oci_cost.sh -d 30 -g DAILY # last 30 days, by service ./scripts/oci_cost.sh -g MONTHLY -d 90 # last 3 months, month-aligned ``

  1. Search the KB before deep debugging:

``bash python3 scripts/kb_lookup.py "usage api cost" cost ``

Read [../../references/cost-management.md](../../references/cost-management.md) for command shapes and [../../references/tenancy-safety.md](../../references/tenancy-safety.md) for the safety rules.

Routing — pick the task

| Request mentions… | Go to | |---|---| | total spend, cost by service, "what am I paying for" | Spend summary | | cost by compartment / region / tag | Grouped usage | | budget status, am I over budget, forecast | Budgets | | set up alerts, guardrails, no budget exists | Guardrail setup → oci-iam-admin | | cost-tracking tag, chargeback, showback | Cost-tracking tags |

Common multi-step flows

| Task | Sequence | |------|----------| | Answer "what are we paying for" | oci_cost.sh -d 30 → read top spend drivers → drill down with --group-by compartmentName / cost-tracking tag | | Investigate a cost spike | oci_cost.sh by service → group-by compartment to localize → audit who created the resource (→ oci-log-analytics Audit query) → recommend a budget alert (→ oci-iam-admin) | | Set up chargeback / showback | define a cost-tracking tag (→ oci-iam-admin) → --group-by-tag usage query → recurring monthly report | | No budget exists yet | oci_cost.sh for current spend + forecast → recommend a budget + 80% alert → defer creation to oci-iam-admin (it gates the mutation) |

Common tasks

# One-shot spend-by-service + budgets (the default view).
./scripts/oci_cost.sh -d 30

# Spend grouped by compartment (Usage API, COST query).
oci_cli usage-api usage-summary request-summarized-usages \
  --tenant-id  \
  --time-usage-started 2026-05-01T00:00:00Z \
  --time-usage-ended   2026-06-01T00:00:00Z \
  --granularity MONTHLY --query-type COST \
  --group-by file://

# Spend grouped by a cost-tracking tag (chargeback/showback).
oci_cli usage-api usage-summary request-summarized-usages \
  --tenant-id  \
  --time-usage-started 2026-05-01T00:00:00Z \
  --time-usage-ended   2026-06-01T00:00:00Z \
  --granularity MONTHLY --query-type COST \
  --group-by-tag file://

# List budgets — NOTE the triple-nested path (service/category/resource).
oci_cli budgets budget budget list --compartment-id  \
  --query 'data[].{name:"display-name",limit:amount,spent:"actual-spend",forecast:"forecasted-spend"}'

# Forecast vs limit — flag budgets trending over.
./scripts/oci_cost.sh -d 30 | awk '/forecast=/ {print}'

Gotchas

  • Budget CLI path is budgets budget budget list — the OCI CLI nests the

service (budgets), category (budget), and resource (budget); budgets budget list does not exist.

  • Usage API needs an explicit --tenant-id. For config auth oci_cost.sh

reads it from ~/.oci/config; for instance/resource-principal auth pass -t or set OCI_SKILLS_TENANCY.

  • Timestamps must align to the granularity. DAILY needs a midnight-UTC

start; MONTHLY needs the first of the month. oci_cost.sh aligns automatically.

  • usage-reports grant required: `allow group to read usage-report in

tenancy`. Without it the Usage API returns nothing (not an error).

  • Latency: cost data lags real time by hours; today's spend is usually $0.

Safety notes

  • Read-only by default. Reporting changes nothing. Creating budgets/alerts is

a mutation — route to oci-iam-admin, which gates it via run_action.

  • Never print OCIDs. The aggregated views are service/compartment names +

amounts by construction; if you query raw items, pipe through redact.

  • Never invent oci flags. Fetch the exact command shape first:

python3 scripts/oci_cli_help.py .

Expected output

**Finding** — spend total + top cost drivers, and budget status (names, not OCIDs).
**Evidence** — redacted oci_cost.sh / Usage API result.
**Action** — read-only by default; guardrail commands (budget + alert) shown but
gated, deferred to oci-iam-admin to execute.
**Verification** — re-run oci_cost.sh showing the budget/alert now present.
**KB** — KB entry used, or new KB- added.

Official documentation

Cost Analysis · Budgets. Full list in the [cost-management reference](../../references/cost-management.md).

Open Knowledge Format grounding — every doc link here is registered and liveness-checked in the [oracle-docs.md index](../../references/oracle-docs.md) (the pack's single source of truth). When extending this skill to build an OCI customer solution, cite the most specific official page through that index so every claim stays verifiable; the non-official MCP gateway is never a source of truth.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet — be the first.

Versions

  • v0.1.0 Imported from the upstream source.