pith. sign in
Pith Number

pith:27AKX2E2

pith:2026:27AKX2E2OHTD3AVNAAZVZW2DE3
not attested not anchored not stored refs resolved

SoK: Agentic Skills -- Beyond Tool Use in LLM Agents

Baihe Ma, Delong Li, Guangsheng Yu, Haiyu Deng, Qin Wang, Xu Wang, Yanna Jiang

Agentic skills function as reusable procedural modules that let LLM agents handle long-horizon tasks reliably across domains.

arxiv:2602.20867 v1 · 2026-02-24 · cs.CR · cs.AI · cs.CE · cs.ET

Add to your LaTeX paper
\usepackage{pith}
\pithnumber{27AKX2E2OHTD3AVNAAZVZW2DE3}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp
2 Internet Archive
3 Author claim open · sign in to claim
4 Citations open
5 Replications open
Portable graph bundle live · download bundle · merged state
The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

Agentic skills are reusable procedural modules with explicit applicability conditions, execution policies, termination criteria, and interfaces that operate reliably across tasks; their adoption introduces supply-chain and prompt-injection risks, as shown by the ClawHavoc campaign in which nearly 1,200 malicious skills infiltrated a major marketplace.

C2weakest assumption

That the two proposed taxonomies (seven design patterns and representation-by-scope) provide a sufficiently complete and stable organizing framework for the rapidly evolving space of agentic skills, and that the security implications drawn from the single ClawHavoc case study generalize to other agent platforms.

C3one line summary

The paper systematizes agentic skills beyond tool use, providing design pattern and representation-scope taxonomies plus security analysis of malicious skill infiltration in agent marketplaces.

References

76 extracted · 76 resolved · 34 Pith anchors

[1] WebArena: A Realistic Web Environment for Building Autonomous Agents 2024 · arXiv:2307.13854
[2] SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering 2024 · arXiv:2405.15793
[3] Measuring and augmenting large language models for solving capture-the-flag challenges, 2025
[4] HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Face 2023 · arXiv:2303.17580
[5] Can large language model agents simulate human trust behavior? 2024

Formal links

3 machine-checked theorem links

Cited by

31 papers in Pith

Receipt and verification
First computed 2026-05-17T23:39:19.876413Z
Builder pith-number-builder-2026-05-17-v1
Signature Pith Ed25519 (pith-v1-2026-05) · public key
Schema pith-number/v1.0

Canonical hash

d7c0abe89a71e63d82ad00335cdb4326fec5ac724461a9ec6c2882233148b40f

Aliases

arxiv: 2602.20867 · arxiv_version: 2602.20867v1 · doi: 10.48550/arxiv.2602.20867 · pith_short_12: 27AKX2E2OHTD · pith_short_16: 27AKX2E2OHTD3AVN · pith_short_8: 27AKX2E2
Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/27AKX2E2OHTD3AVNAAZVZW2DE3 \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: d7c0abe89a71e63d82ad00335cdb4326fec5ac724461a9ec6c2882233148b40f
Canonical record JSON
{
  "metadata": {
    "abstract_canon_sha256": "b2538c823619093b7d52d31cf22ebe645647f5d5a86d6344fe246b28a40d5b7e",
    "cross_cats_sorted": [
      "cs.AI",
      "cs.CE",
      "cs.ET"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.CR",
    "submitted_at": "2026-02-24T13:11:38Z",
    "title_canon_sha256": "8586e819e4dabd6f63deafc404829d9d39f16b79aae0d6894ffc8a81e13ce8f1"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2602.20867",
    "kind": "arxiv",
    "version": 1
  }
}