Ready to supercharge your AI agents?

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

Get Started Free →

Repository agent instructions (entry) · asymmetric-al | DecimalAI

asymmetric-al/repo-entry

SkillScore82Maintained

Pointer to the canonical agent instruction and skill system for this monorepo

agent-toolingby asymmetric-al

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

Coding pack

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

pip install decimalai

2 · Get the skill — writes .claude/skills/asymmetric-al-repo-entry/SKILL.md

decimalai skills pull asymmetric-al-repo-entry --out .claude/skills/

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

curl -fsSL --create-dirs -o .claude/skills/asymmetric-al-repo-entry/SKILL.md https://app.decimal.ai/s/asymmetric-al-repo-entry/SKILL.md

Raw file: /s/asymmetric-al-repo-entry/SKILL.md ↗ · machine-readable summary at /s/asymmetric-al-repo-entry.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

22 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 25d ago

—-28% tokens0% turns82% of whole cases pass with skill (18/22)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-17	✗→✓	▲ Improved	-39%	0%
case-04	✗→✓	▲ Improved	-26%	0%
case-05	✗→✓	▲ Improved	-30%	0%
case-06	✗→✓	▲ Improved	-47%	0%
case-07	✗→✓	▲ Improved	-65%	0%

Showing 5 of 22 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v5 — 8/26/2026

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

asymmetric-al / core

SKILL.md

View on GitHub ↗

License AGPL-3.0Imported 8/26/2026Version v3

Skill contents

SKILL.md

Repository agent instructions (entry)

This skill is the maintenance entrypoint for the instruction and skill system. Root AGENTS.md already provides ordinary repository orientation.

When to use

You are auditing or changing AGENTS.md, rule/skill routing, or instruction budgets.
A tool or workflow only surfaced a root SKILL.md and you need the full picture.
You changed or added something under docs/ai/skills/ and must refresh mirrors.

Workflow

Read and follow root AGENTS.md, the nearest scoped instructions, and

openspec/specs/agent-instruction-system/spec.md.

For canonical task skills maintained by this repo, use docs/ai/skills/*/SKILL.md and docs/ai/rules/agent-skill-routing.md.
After changing skills under docs/ai/skills/, run bun run skills:sync and ensure bun run skills:verify passes before committing.
Optional tool-specific or ecosystem skills may originate under

.agents/skills/ and be mirrored into .cursor/skills/ and .claude/skills/; use them only when their SKILL.md description matches and they remain subordinate to canonical repo guidance.

Checklist

] Root and scoped instruction sources opened for the maintenance task.
] Relevant docs/ai/skills/<topic>/SKILL.md files loaded per docs/ai/rules/agent-skill-routing.md.
] If docs/ai/skills/ was edited: bun run skills:sync and bun run skills:verify run and clean.

See also

Discover / install skills: docs/ai/skills/find-skills/SKILL.md
Instruction contract: openspec/specs/agent-instruction-system/spec.md

Illustrative Want to see what asymmetric-al/repo-entry would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

8/27/2026, 04:32 UTC · 22 cases

Measured on 18 comparable cases — below the 22 needed to publish a lift, so this result is shown here but not on the registry card or leaderboard.

Lift

+64%

82% vs 18%

Agent turns

0%

1.0 vs 1.0

Tokens

-28%

1,190.5 vs 1,655.11

fail → pass · 14pass → pass · 4fail → fail · 4

Per-case results (22)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-17 Which file path holds the formal open specification for the agent instruction system in this codebase?	fail→pass	10,248	2,845	-72%	1	1	0%	1,472	901	-39%	0	0	—
case-01 I've just added a new canonical skill definition to the AI skills folder. Can you follow our repository's instruction maintenance workflow to sync any generated mirrors, run whatever checks are required, and verify that the changes are valid before I commit?	fail→fail	4,348	2,588	-40%	1	1	0%	206	849	+312%	0	0	—
case-02 We need to audit our top-level agent instructions and skill routing rules because some tools are only picking up the root skill file. Could you inspect the repository's instruction system spec, review AGENTS.md and our routing docs, and provide a summary checklist of any necessary updates?	fail→fail	4,517	6,027	+33%	1	1	0%	227	930	+310%	0	0	—
case-03 I made some edits to our agent routing configurations under the docs directory. Please go through the standard repository maintenance steps for AI instructions—check the relevant spec files, execute any required synchronization build scripts, and give me a status report on whether all checks pass.	fail→fail	3,784	6,323	+67%	1	1	0%	279	1,010	+262%	0	0	—
case-04 We updated the canonical code review skill under docs/ai/skills/code-review/SKILL.md using our custom package runner scripts. I usually just run npm run build directly. How should I update and validate the mirrors before pushing to main?	fail→pass	10,181	5,288	-48%	1	1	0%	1,802	1,330	-26%	0	0	—
case-05 I am adding a core, repo-wide task skill for database migrations that every developer on this project must follow. Should I create this canonical skill under .cursor/skills/db-migrations/SKILL.md or directly under .agents/skills/db-migrations/SKILL.md?	fail→pass	12,656	5,380	-57%	1	1	0%	2,270	1,585	-30%	0	0	—
case-06 When optional editor-specific agent workflows are added to the repo's generic agent skills folder, which IDE mirror directories should they be synchronized into to ensure editor compatibility?	fail→pass	9,500	1,984	-79%	1	1	0%	1,550	817	-47%	0	0	—
case-07 Our agent instruction token budget seems bloated. I plan to trim AGENTS.md and refactor routing rules. Which specific specification document defines the contract and rules governing this agent instruction architecture?	fail→pass	16,721	3,214	-81%	1	1	0%	2,665	938	-65%	0	0	—
case-08 I am writing a CI validation step that checks if generated AI skill mirrors match their source definitions under docs/ai/skills/. Should the pipeline invoke npm run verify-skills or yarn skills:check?	fail→pass	8,494	4,831	-43%	1	1	0%	1,592	1,338	-16%	0	0	—
case-09 Where in the repository documentation hierarchy are skill routing decisions and rule mappings explicitly defined for AI agents?	fail→pass	12,260	5,170	-58%	1	1	0%	2,145	1,371	-36%	0	0	—
case-10 My editor tool only surfaced a single root SKILL.md file when opening the workspace. How do I navigate to find the complete agent architecture instructions?	fail→pass	11,126	4,648	-58%	1	1	0%	1,990	1,265	-36%	0	0	—
case-11 If a tool-specific skill rule in .cursor/skills/ conflict with a instruction in docs/ai/skills/, which instruction takes precedence during agent execution?	fail→pass	7,833	4,007	-49%	1	1	0%	1,272	1,199	-6%	0	0	—
case-16 Under what condition should an agent adopt an optional ecosystem skill located in .agents/skills/?	fail→pass	7,558	5,335	-29%	1	1	0%	1,187	1,410	+19%	0	0	—
case-12 We need to run our skill synchronization utility in an automated environment. What javascript runtime package runner is configured to execute these scripts in our project?	fail→pass	7,906	2,412	-69%	1	1	0%	1,388	828	-40%	0	0	—
case-13 A new contributor wants to discover and install existing skills maintained in this repository. Which documentation file provides skill discovery instructions?	fail→pass	8,412	2,601	-69%	1	1	0%	1,383	930	-33%	0	0	—
case-14 Before submitting a PR that edits AI agent guidelines, what validation commands must run and report clean status?	fail→pass	9,821	5,744	-42%	1	1	0%	1,659	934	-44%	0	0	—
case-15 When working on subfolder-specific features, what instruction files must be read alongside the root AGENTS.md to ensure compliance?	pass→pass	8,327	6,390	-23%	1	1	0%	1,514	1,697	+12%	0	0	—
case-18 What primary purpose does the root AGENTS.md file serve for AI agents entering the repository workspace?	pass→pass	7,452	3,307	-56%	1	1	0%	1,224	976	-20%	0	0	—
case-19 Does modifying a file under docs/ai/skills/ testing/ require running the sync script, or only edits to AGENTS.md?	pass→pass	8,739	3,738	-57%	1	1	0%	1,479	1,115	-25%	0	0	—
case-20 I am refactoring a TypeScript backend service under src/services/user.ts to use async/await instead of callbacks. What command should I run to sync agent skills for this code change?	fail→pass	12,839	4,998	-61%	1	1	0%	2,032	1,216	-40%	0	0	—
case-21 What is our team's standard git branch naming strategy for bugfixes and release tags when pushing code to GitHub?	fail→fail	5,281	5,390	+2%	1	1	0%	948	726	-23%	0	0	—
case-22 I need to add a new npm package dependency to package.json. Should I run bun add or edit package.json manually?	pass→pass	6,155	6,319	+3%	1	1	0%	1,168	1,579	+35%	0	0	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 18 counted toward the lift figure. The other 4 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +64 percentage points is the difference between those two pass rates over the 18 comparable cases.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 18.2%
Pass rate with the skill: 81.8%
Cases attempted: 22
Cases counted toward lift: 18
Cases that regressed: 0
Skill version measured: v3
Run completed: 8/27/2026, 04:32 UTC

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-17

Prompt

Which file path holds the formal open specification for the agent instruction system in this codebase?

Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.

Show both answers

SkillScore blends 4 signals1 of 4 complete

BenchmarkPass rate WITH skill on a verified run82

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns0% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

Backed by 1 of 4 signals — each added signal makes the score more robust.

Verified on gemini-3.6-flash · re-tested 24 days ago

Model	Method	Date	Lift
gemini-3.6-flash	verified	8/20/2026	+74%

Other measured skills in the registry, with their headline benchmark lift.