Pith Number

pith:7KZP4YJ2

pith:2026:7KZP4YJ2IFUSQ4GB7ZHOEDQJ7W

not attested not anchored not stored refs resolved

Is Grep All You Need? How Agent Harnesses Reshape Agentic Search

Akhil Kasturi, Anmol Gulati, Elias Lumer, Sahil Sen, Vamse Kumar Subbiah

Grep retrieval often beats vector search for accuracy in LLM agent workflows, though harness and tool-calling style drive most of the performance difference.

arxiv:2605.15184 v1 · 2026-05-14 · cs.CL

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{7KZP4YJ2IFUSQ4GB7ZHOEDQJ7W}

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

Across Chronos and the provider CLIs, grep generally yields higher accuracy than vector retrieval in our comparisons in experiment 1; at the same time, overall scores still depend strongly on which harness and tool-calling style is used, even when the underlying conversation data are the same.

C2weakest assumption

That the 116-question sample from LongMemEval and the chosen harness implementations (Chronos, Claude Code, Codex, Gemini CLI) are representative of broader agentic search performance.

C3one line summary

Grep retrieval generally outperforms vector retrieval in agentic search tasks, with performance varying strongly by agent harness and tool-calling style.

References

32 extracted · 32 resolved · 8 Pith anchors

[1] Akari Asai, Zeqiu Wu, Yizhong Wang, Avirup Sil, and Hannaneh Hajishirzi. 2024. Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection. InProceedings of ICLR 2024

[2] Evaluating Large Language Models Trained on Code 2021 · arXiv:2107.03374

[3] Gordon V. Cormack, Charles L. A. Clarke, and Stefan Buettcher. 2009. Reciprocal Rank Fusion Outperforms Condorcet and Individual Rank Learning Methods. In Proceedings of SIGIR. 758–759 2009

[4] Thibault Formal, Carlos Lassance, Benjamin Piwowarski, and Stéphane Clinchant

[5] doi:10.48550/ARXIV.2109.10086 2021

Receipt and verification

First computed	2026-05-17T21:40:25.119335Z
Last reissued	2026-05-17T21:57:18.501018Z
Builder	pith-number-builder-2026-05-17-v1
Signature	unsigned_v0
Schema	pith-number/v1.0

Canonical hash

fab2fe613a41692870c1fe4ee20e09fd92ce5abeb23353b0b3c87fc858865ec5

Aliases

arxiv: 2605.15184 · arxiv_version: 2605.15184v1 · pith_short_12: 7KZP4YJ2IFUS · pith_short_16: 7KZP4YJ2IFUSQ4GB · pith_short_8: 7KZP4YJ2

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/7KZP4YJ2IFUSQ4GB7ZHOEDQJ7W \
  | 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: fab2fe613a41692870c1fe4ee20e09fd92ce5abeb23353b0b3c87fc858865ec5

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "06b18eb3db10ed5898a0f5b5a6f0b616d238d315cff6419f208ce59c3287fabf",
    "cross_cats_sorted": [],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.CL",
    "submitted_at": "2026-05-14T17:58:41Z",
    "title_canon_sha256": "882523d375a459c9c109456a0bbf0fb20ee9b760d5292c9be0a2ac2043cbfcee"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.15184",
    "kind": "arxiv",
    "version": 1
  }
}