{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:LZJMXACJ2QOP7HZXTFKINNIWCF","short_pith_number":"pith:LZJMXACJ","schema_version":"1.0","canonical_sha256":"5e52cb8049d41cff9f37995486b516115c78fdf867d1b9c76f5964c7ca721b64","source":{"kind":"arxiv","id":"2604.11088","version":2},"attestation_state":"computed","paper":{"title":"Guardrails Beat Guidance: A Large-Scale Study of Rules, Skills, and Persistent Configuration for Coding Agents","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"Negative constraints raise AI coding agent success rates while positive instructions lower them, and random rules match expert ones.","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Bing Zhu, Guanghui Wang, Peiyang He, Wei Qiu, Xing Zhang, Yanwei Cui, Ziyuan Li","submitted_at":"2026-04-13T07:10:01Z","abstract_excerpt":"Random rules improve a coding agent's task performance as much as expert-curated ones (both $+13.8$pp on a discriminative subset of SWE-bench Verified), and in our data every individually beneficial rule is a negative constraint (\"do not refactor unrelated code\"), while every individually harmful one is a positive directive (\"follow code style\"). We arrive at these findings through the first large-scale controlled study of agent rule files (\\texttt{CLAUDE.md}, \\texttt{.cursorrules}, and the broader family of agent skills, plugin manifests, and persona definitions): we scrape 679 rule files (25"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2604.11088","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-04-13T07:10:01Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"9e767301e3cdcde0e6a2b375adcccddcc3020265586dd650c38257304c39e7ca","abstract_canon_sha256":"9ddbe9ead40d1e75fbf7d5c1bd1d2688dc8c2b4a350c36f98988ecc1294a6228"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-29T01:05:09.361213Z","signature_b64":"frFbqo4DjE5cPvigwi9VrOaBCuTF3ZA+9Mumiw1j7LjhRmjGjN+xbqtrOtfSY747pBGGJl/VnYaJpqrFCBhODw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5e52cb8049d41cff9f37995486b516115c78fdf867d1b9c76f5964c7ca721b64","last_reissued_at":"2026-05-29T01:05:09.360267Z","signature_status":"signed_v1","first_computed_at":"2026-05-29T01:05:09.360267Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Guardrails Beat Guidance: A Large-Scale Study of Rules, Skills, and Persistent Configuration for Coding Agents","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"Negative constraints raise AI coding agent success rates while positive instructions lower them, and random rules match expert ones.","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Bing Zhu, Guanghui Wang, Peiyang He, Wei Qiu, Xing Zhang, Yanwei Cui, Ziyuan Li","submitted_at":"2026-04-13T07:10:01Z","abstract_excerpt":"Random rules improve a coding agent's task performance as much as expert-curated ones (both $+13.8$pp on a discriminative subset of SWE-bench Verified), and in our data every individually beneficial rule is a negative constraint (\"do not refactor unrelated code\"), while every individually harmful one is a positive directive (\"follow code style\"). We arrive at these findings through the first large-scale controlled study of agent rule files (\\texttt{CLAUDE.md}, \\texttt{.cursorrules}, and the broader family of agent skills, plugin manifests, and persona definitions): we scrape 679 rule files (25"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Rules improve performance by 7--14 percentage points, but random rules help as much as expert-curated ones -- suggesting rules work through context priming rather than specific instruction. Negative constraints are the only individually beneficial rule type, while positive directives actively hurt.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That differences in agent success rates are caused by the rule types themselves rather than unmeasured factors in how rules are inserted into prompts or variations across agent runs.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Negative constraints in agent rules improve coding performance via context priming while positive directives degrade it, with collective rules remaining helpful up to 50 rules.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Negative constraints raise AI coding agent success rates while positive instructions lower them, and random rules match expert ones.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"e7dd00151feaf1be5a651358c4b6a755fc576a6d4ca45b0f75e2f20f231d11e7"},"source":{"id":"2604.11088","kind":"arxiv","version":2},"verdict":{"id":"a2e26d41-360e-4016-8608-abd8319657b3","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T15:33:36.613027Z","strongest_claim":"Rules improve performance by 7--14 percentage points, but random rules help as much as expert-curated ones -- suggesting rules work through context priming rather than specific instruction. Negative constraints are the only individually beneficial rule type, while positive directives actively hurt.","one_line_summary":"Negative constraints in agent rules improve coding performance via context priming while positive directives degrade it, with collective rules remaining helpful up to 50 rules.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That differences in agent success rates are caused by the rule types themselves rather than unmeasured factors in how rules are inserted into prompts or variations across agent runs.","pith_extraction_headline":"Negative constraints raise AI coding agent success rates while positive instructions lower them, and random rules match expert ones."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.11088/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2604.11088","created_at":"2026-05-29T01:05:09.360412+00:00"},{"alias_kind":"arxiv_version","alias_value":"2604.11088v2","created_at":"2026-05-29T01:05:09.360412+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.11088","created_at":"2026-05-29T01:05:09.360412+00:00"},{"alias_kind":"pith_short_12","alias_value":"LZJMXACJ2QOP","created_at":"2026-05-29T01:05:09.360412+00:00"},{"alias_kind":"pith_short_16","alias_value":"LZJMXACJ2QOP7HZX","created_at":"2026-05-29T01:05:09.360412+00:00"},{"alias_kind":"pith_short_8","alias_value":"LZJMXACJ","created_at":"2026-05-29T01:05:09.360412+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2605.22148","citing_title":"Ratchet: A Minimal Hygiene Recipe for Self-Evolving LLM Agents","ref_index":6,"is_internal_anchor":true},{"citing_arxiv_id":"2605.19576","citing_title":"Library Drift: Diagnosing and Fixing a Silent Failure Mode in Self-Evolving LLM Skill Libraries","ref_index":17,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LZJMXACJ2QOP7HZXTFKINNIWCF","json":"https://pith.science/pith/LZJMXACJ2QOP7HZXTFKINNIWCF.json","graph_json":"https://pith.science/api/pith-number/LZJMXACJ2QOP7HZXTFKINNIWCF/graph.json","events_json":"https://pith.science/api/pith-number/LZJMXACJ2QOP7HZXTFKINNIWCF/events.json","paper":"https://pith.science/paper/LZJMXACJ"},"agent_actions":{"view_html":"https://pith.science/pith/LZJMXACJ2QOP7HZXTFKINNIWCF","download_json":"https://pith.science/pith/LZJMXACJ2QOP7HZXTFKINNIWCF.json","view_paper":"https://pith.science/paper/LZJMXACJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2604.11088&json=true","fetch_graph":"https://pith.science/api/pith-number/LZJMXACJ2QOP7HZXTFKINNIWCF/graph.json","fetch_events":"https://pith.science/api/pith-number/LZJMXACJ2QOP7HZXTFKINNIWCF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LZJMXACJ2QOP7HZXTFKINNIWCF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LZJMXACJ2QOP7HZXTFKINNIWCF/action/storage_attestation","attest_author":"https://pith.science/pith/LZJMXACJ2QOP7HZXTFKINNIWCF/action/author_attestation","sign_citation":"https://pith.science/pith/LZJMXACJ2QOP7HZXTFKINNIWCF/action/citation_signature","submit_replication":"https://pith.science/pith/LZJMXACJ2QOP7HZXTFKINNIWCF/action/replication_record"}},"created_at":"2026-05-29T01:05:09.360412+00:00","updated_at":"2026-05-29T01:05:09.360412+00:00"}