Skills context-degradation

📦

context-degradation

Name: context-degradation
Author: Asmayaseen

Content revision r2 Safe ⚙️ External commands

Diagnose Context Degradation in AI Agents

Long contexts can make AI agents overlook facts, repeat errors, or apply conflicting instructions. This skill identifies common degradation patterns and recommends focused mitigation strategies.

Supports: Claude Codex Code(CC)

🥈 80 Silver

Install with my Agent

Copy this request to your Agent. It includes the canonical Skill page and manifest.

Agent request

Review the Skillstore skill "context-degradation" from https://skillstore.io/skills/asmayaseen-context-degradation.md and its manifest at https://skillstore.io/api/skills/asmayaseen-context-degradation/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.

Your Agent should still show its plan and request any confirmation required by the security policy.

Agent-readable resources

Use these links when an AI agent, crawler, or script needs clean context instead of reading the full page.

Markdown detail GET /skills/asmayaseen-context-degradation.md Signed manifest GET /api/skills/asmayaseen-context-degradation/manifest Signed lockfile GET /api/skills/asmayaseen-context-degradation/lockfile

Test it

Using "context-degradation". A support agent ignores a refund policy placed halfway through a very long context.

Expected outcome:

Likely pattern: lost in middle. Move the policy near the current task, repeat its key constraint at the end, and test recall by position.

Using "context-degradation". An agent repeats an incorrect customer status after a failed tool response was summarized as successful.

Expected outcome:

Likely pattern: context poisoning. Remove the incorrect summary, reload verified tool output, and label the correction before continuing.

Using "context-degradation". A research agent combines current guidance with an older document that recommends the opposite action.

Expected outcome:

Likely pattern: context clash. Mark the conflict, apply source and version priority, and exclude the outdated document.

Security Audit

Safe

v9 • 7/23/2026 Open versioned report

All five static findings are false positives caused by Markdown fences, ordinary explanatory prose, and a literal heuristic phrase. No external command execution, system reconnaissance, prompt injection, data exfiltration, or other semantic threat was found.

Files scanned

996

Lines analyzed

Review items

False positives ignored

Risk Factors

⚙️ External commands (3)

SKILL.md:162-169 SKILL.md:169-172 SKILL.md:172-188

No confirmed security findings were detected by the latest completed static and semantic audit. This does not prove the skill has no side effects.

Audited by: codex View Audit History →

Share & cite this report

Share the versioned assessment report, neutral badge, embed card, and citations. Skillstore reports evidence without deciding whether this Skill is safe.

Open versioned report

Copy report link

https://skillstore.io/skills/asmayaseen-context-degradation/audits/9?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Markdown badge

[![Skillstore security assessment](https://skillstore.io/badges/skills/asmayaseen-context-degradation/security.svg)](https://skillstore.io/skills/asmayaseen-context-degradation?utm_source=security_passport_badge)

HTML badge

<a href="https://skillstore.io/skills/asmayaseen-context-degradation?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/asmayaseen-context-degradation/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

Embed card

<iframe src="https://skillstore.io/embed/skills/asmayaseen-context-degradation.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>

Academic citations (APA · BibTeX · CFF)

APA citation

Asmayaseen. (2026). context-degradation security audit report (audit version 9) [Author version unspecified]. Skillstore. https://skillstore.io/skills/asmayaseen-context-degradation/audits/9

BibTeX citation

@techreport{asmayaseen-asmayaseen-context-degradation-2026,
  author = {Asmayaseen},
  title = {context-degradation security audit report (audit version 9)},
  institution = {Skillstore},
  year = {2026},
  number = {9},
  url = {https://skillstore.io/skills/asmayaseen-context-degradation/audits/9},
  note = {Author version unspecified}
}

CITATION.cff

cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "context-degradation security audit report (audit version 9)"
version: "unspecified"
type: report
authors:
  - name: "Asmayaseen"
date-released: "2026-07-23"
url: "https://skillstore.io/skills/asmayaseen-context-degradation/audits/9"
identifiers:
  - type: other
    value: "skillstore:asmayaseen-context-degradation:audit:9"
    description: "Skillstore immutable audit report identifier"

Compare variants

4 installable variants

Each author remains a separate installable skill. The recommended variant is ranked by Skillstore evidence.

Why this variant is first

Highest Skillstore Score

Author / source Skillstore Score Evidence Confidence Skillstore usage Updated

ChakshuGautam Recommended

chakshugautam-context-degradation

Skillstore Score 81

Evidence Confidence High

Skillstore usage 15

Updated

2026-08-21

muratcankoylan

muratcankoylan-context-degradation

Skillstore Score 80

Evidence Confidence High

Skillstore usage 21

Updated

2026-08-21

Asmayaseen Current

asmayaseen-context-degradation

Skillstore Score 80

Evidence Confidence High

Skillstore usage 15

Updated

2026-08-21

sickn33

sickn33-context-degradation

Skillstore Score 78

Evidence Confidence High

Skillstore usage 16

Updated

2026-08-21

Skillstore Score

Why this score Evidence Confidence: High

Architecture

Maintainability

Content

Community

Spec Compliance

What You Can Build

Debug Long Agent Sessions

Identify why an agent starts ignoring goals or selecting incorrect tools as its conversation grows.

Design Context Controls

Choose compaction, filtering, masking, or isolation controls for production agent workflows.

Plan Context Evaluations

Define tests that vary context length, information position, distractors, and conflicting sources.

Try These Prompts

Identify a Degradation Pattern

Review this agent failure: [failure description]. Identify the most likely context degradation pattern and explain the evidence.

Recommend a Mitigation

Analyze this context workflow: [workflow]. Recommend one immediate mitigation and one architectural improvement. Explain which degradation risk each change addresses.

Audit Context Structure

Audit this context layout: [layout]. Assess critical information placement, irrelevant content, conflicting sources, and accumulated errors. Prioritize the findings.

Design a Degradation Evaluation

Design an evaluation for [agent system]. Vary context length, evidence position, distractor count, and source conflicts. Define metrics, thresholds, and recovery tests.

Best Practices

Measure task quality across increasing context lengths instead of relying on window size alone.
Keep goals, current constraints, and verified findings near attention-favored positions.
Validate retrieved content and preserve source versions before adding it to context.