Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Get Started Free →

Deep Reference · agent-sh | DecimalAI

Registry skillby agent-sh

Deep Reference

agent-sh/deep-reference

Source agent-sh/agnix ↗License Apache-2.0

Use when validating deep references

SkillScore50Maintained

Automatic

Doc Ops pack

Last verified 8/3/2026

What it changes

Passed 11 of 22 test cases with the skill, and 7 of 22 without it.

Without skill32%

With skill50%

Checks it passed that the model alone missed

Includes `API Gateway` as a core system architecture component.
Requires Prometheus counters to use the `_total` metric name suffix.
Requires `additionalProperties: false` on all OpenAPI request object schemas.

and 2 more in the full report

5 fixed6 passed either way10 still failing1 broke

Pass-rate lift: +18 pp
Tokens: +97%
Agent turns: 0%
Response time: +86%

Model gemini-3.6-flashMeasured 8/3/2026v111 of 22 cases passed with skill

Tokens, turns and time describe resource use, not a dollar price.

Use this skill

Save it into your project's skills folder. Claude Code picks it up in its next session.

curl -fsSL --create-dirs -o .claude/skills/agent-sh-deep-reference/SKILL.md https://app.decimal.ai/s/agent-sh-deep-reference/SKILL.md

Want it pinned and updatable? pip install "decimalai>=0.15", then decimalai skills add agent-sh-deep-reference, and start Claude Code with claude --plugin-dir "$(git rev-parse --show-toplevel)/.agents"

1 · One-time — install the CLI (docs ↗)

pip install "decimalai>=0.15"

2 · Get the skill — writes .agents/skills/agent-sh-deep-reference/ and pins it in .decimal/skills.lock

decimalai skills add agent-sh-deep-reference

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/agent-sh-deep-reference/SKILL.md https://app.decimal.ai/s/agent-sh-deep-reference/SKILL.md

Raw file: /s/agent-sh-deep-reference/SKILL.md ↗ · machine-readable summary at /s/agent-sh-deep-reference.json ↗

Safety checks: 3 of 3 passed

Code scan CleanIntent PassedContent Clean

See it in action

One of 5 test cases the skill fixed. Same model, same prompt, with and without the skill.

1 of 5

Prompt

We are implementing OAuth2 PKCE in our authentication service. Rather than giving me standard RFC 7636 steps, check our nested documentation tree for our organization's token exchange rules and summarize the flow.

What was checked, without and with the skill
Checked	Without skill	With skill
Specifies `S256` as the required `code_challenge_method` parameter for PKCE token exchanges.	Missed	Passed

case-04 · gemini-3.6-flash · 8/3/2026 · answers are capped and redacted

What's inside

The instructions your model reads when the skill loads.

SKILL.md1 linessingle file

See references/deep/nested/guide.md for details.

Track record

Benchmark

11 / 22

test cases passed, vs 7 / 22 without

Adoption

Users: 0
Times used: 0

Safety and source details

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/1/2026

View 15 scanner checks

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

agent-sh / agnix

tests/fixtures/skills/deep-reference/SKILL.md

View on GitHub ↗

License Apache-2.0Imported 8/1/2026Version v1

More skills from the same source and nearby in the registry, with their headline benchmark lift.