aiskillstore/marketplace / skills/89jobrien/context-management
context-management
VerifiedContext management specialist for multi-agent workflows and long-running tasks. Use when coordinating multiple agents, preserving context across sessions, or managing complex project state. Focuses on context compression and distribution.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 2.5 KB |
| memory-management.md | markdown | 7.9 KB |
| references/patterns.md | markdown | 2.4 KB |
| skill-report.json | config | 23.7 KB |
skill-report.json
View on GitHub{
"schema_version": "2.0",
"meta": {
"generated_at": "2026-07-05T11:36:30.822Z",
"slug": "89jobrien-context-management",
"source_url": "https://github.com/89jobrien/steve/tree/main/steve/skills/context-management",
"source_ref": "main",
"model": "codex",
"analysis_version": "3.0.0",
"source_type": "community",
"content_hash": "2eccfcb02e5b18c902adce0845ef7befedd082abb354efc1e4455e6bb995a020",
"tree_hash": "9c2b8ccb85caf6dbb8cac4404dfe1a560239657bc93ce1dfd2ecca57c3f7299c",
"provenance": {
"model_effective": "codex",
"fallback_chain": [
{
"agent": "codex",
"outcome": "success"
}
],
"analysis_version": "3.0.0"
}
},
"skill": {
"name": "context-management",
"description": "Manage context for multi-agent workflows, long-running projects, and persistent memory. The skill helps capture decisions, prepare handoffs, compress context, and organize project knowledge.",
"summary": "Capture, compress, and hand off project context across agents and sessions.",
"icon": "📦",
"version": "1.0.2",
"author": "Joseph OBrien",
"license": "MIT",
"category": "productivity",
"tags": [
"context-management",
"multi-agent",
"handoffs",
"memory",
"workflow"
],
"supported_tools": [
"claude",
"codex",
"claude-code"
],
"risk_factors": [
"external_commands",
"filesystem"
]
},
"security_audit": {
"risk_level": "medium",
"is_blocked": false,
"safe_to_publish": true,
"analysis_status": "ok",
"summary": "Most static findings are false positives caused by markdown code fences or ordinary prose. The confirmed static risks are filesystem examples that use predictable /tmp paths and broad glob matching. A semantic issue remains because the memory guidance suggests preserving credentials in persistent context.",
"remediation": [
{
"issue": "Insecure temporary artifact pattern",
"suggestion": "Use a dedicated workspace directory, randomized filenames, strict permissions, and cleanup instead of predictable /tmp files and broad globs.",
"severity": "medium"
},
{
"issue": "Credential storage guidance",
"suggestion": "Remove credentials from the preservation list and instruct users to store only redacted references or secret-manager pointers.",
"severity": "medium"
}
],
"risk_factor_evidence": [
{
"factor": "external_commands",
"evidence": [
{
"file": "references/patterns.md",
"line_start": 105,
"line_end": 113
},
{
"file": "SKILL.md",
"line_start": 40,
"line_end": 42
},
{
"file": "SKILL.md",
"line_start": 42,
"line_end": 44
},
{
"file": "SKILL.md",
"line_start": 44,
"line_end": 46
},
{
"file": "SKILL.md",
"line_start": 46,
"line_end": 50
},
{
"file": "SKILL.md",
"line_start": 50,
"line_end": 52
},
{
"file": "SKILL.md",
"line_start": 52,
"line_end": 88
}
]
},
{
"factor": "filesystem",
"evidence": [
{
"file": "references/patterns.md",
"line_start": 50,
"line_end": 50
},
{
"file": "references/patterns.md",
"line_start": 53,
"line_end": 53
},
{
"file": "references/patterns.md",
"line_start": 53,
"line_end": 53
}
]
}
],
"static_findings": [
{
"id": "blocker:memory-management.md:172:system-reconnaissance",
"category": "blocker",
"severity": "low",
"pattern": "System reconnaissance",
"file": "memory-management.md",
"line_start": 172,
"line_end": 172,
"snippet": "1. Check entity name format (avoid special characters)"
},
{
"id": "blocker:memory-management.md:174:system-reconnaissance",
"category": "blocker",
"severity": "low",
"pattern": "System reconnaissance",
"file": "memory-management.md",
"line_start": 174,
"line_end": 175,
"snippet": "3. Confirm observation format is valid"
},
{
"id": "blocker:memory-management.md:186:system-reconnaissance",
"category": "blocker",
"severity": "low",
"pattern": "System reconnaissance",
"file": "memory-management.md",
"line_start": 186,
"line_end": 186,
"snippet": "- Avoid special characters"
},
{
"id": "blocker:memory-management.md:201:system-reconnaissance",
"category": "blocker",
"severity": "low",
"pattern": "System reconnaissance",
"file": "memory-management.md",
"line_start": 201,
"line_end": 201,
"snippet": "- Avoid circular dependencies"
},
{
"id": "blocker:memory-management.md:272:system-reconnaissance",
"category": "blocker",
"severity": "low",
"pattern": "System reconnaissance",
"file": "memory-management.md",
"line_start": 272,
"line_end": 274,
"snippet": "## Common Pitfalls to Avoid"
},
{
"id": "external_commands:references/patterns.md:105:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "references/patterns.md",
"line_start": 105,
"line_end": 113,
"snippet": "- `path/to/file.ts` - [description]"
},
{
"id": "filesystem:references/patterns.md:50:temp-directory-access",
"category": "filesystem",
"severity": "medium",
"pattern": "Temp directory access",
"file": "references/patterns.md",
"line_start": 50,
"line_end": 50,
"snippet": "output_path = f\"/tmp/agent_{task_id}_{timestamp}.md\""
},
{
"id": "filesystem:references/patterns.md:53:temp-directory-access",
"category": "filesystem",
"severity": "medium",
"pattern": "Temp directory access",
"file": "references/patterns.md",
"line_start": 53,
"line_end": 53,
"snippet": "artifacts = glob.glob(\"/tmp/agent_*.md\")"
},
{
"id": "filesystem:references/patterns.md:53:python-glob-pattern-matching",
"category": "filesystem",
"severity": "low",
"pattern": "Python glob/pattern matching",
"file": "references/patterns.md",
"line_start": 53,
"line_end": 53,
"snippet": "artifacts = glob.glob(\"/tmp/agent_*.md\")"
},
{
"id": "blocker:references/patterns.md:31:network-reconnaissance",
"category": "blocker",
"severity": "low",
"pattern": "Network reconnaissance",
"file": "references/patterns.md",
"line_start": 31,
"line_end": 32,
"snippet": "- What's in progress"
},
{
"id": "external_commands:SKILL.md:40:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 40,
"line_end": 42,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:42:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 42,
"line_end": 44,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:44:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 44,
"line_end": 46,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:46:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 46,
"line_end": 50,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:50:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 50,
"line_end": 52,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:52:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 52,
"line_end": 88,
"snippet": "```"
}
],
"finding_verdicts": [
{
"id": "blocker:memory-management.md:172:system-reconnaissance",
"verdict": "false_positive",
"confidence": 0.96,
"reason": "This is troubleshooting advice about validating entity names. It does not inspect the host system or request reconnaissance."
},
{
"id": "blocker:memory-management.md:174:system-reconnaissance",
"verdict": "false_positive",
"confidence": 0.96,
"reason": "This line asks users to confirm observation format during memory error recovery. It is not system probing or collection."
},
{
"id": "blocker:memory-management.md:186:system-reconnaissance",
"verdict": "false_positive",
"confidence": 0.97,
"reason": "The text advises avoiding special characters in entity names. It contains no reconnaissance behavior or host inspection."
},
{
"id": "blocker:memory-management.md:201:system-reconnaissance",
"verdict": "false_positive",
"confidence": 0.97,
"reason": "The line discusses avoiding circular dependencies in memory relationships. It is conceptual data-model guidance, not system reconnaissance."
},
{
"id": "blocker:memory-management.md:272:system-reconnaissance",
"verdict": "false_positive",
"confidence": 0.98,
"reason": "The section title introduces common memory-management pitfalls. It does not instruct reconnaissance or environment discovery."
},
{
"id": "external_commands:references/patterns.md:105:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.97,
"reason": "The backticks are markdown formatting for a placeholder file path in a handoff template. There is no shell execution context."
},
{
"id": "filesystem:references/patterns.md:50:temp-directory-access",
"verdict": "confirmed",
"confidence": 0.68,
"reason": "The example tells agents to write artifacts under predictable /tmp paths. That can collide with other files or expose coordination data on shared systems.",
"severity": "medium"
},
{
"id": "filesystem:references/patterns.md:53:temp-directory-access",
"verdict": "confirmed",
"confidence": 0.72,
"reason": "The example reads every matching /tmp artifact. Broad temp-directory reads can ingest stale or attacker-created files.",
"severity": "medium"
},
{
"id": "filesystem:references/patterns.md:53:python-glob-pattern-matching",
"verdict": "confirmed",
"confidence": 0.7,
"reason": "The glob pattern supports the same broad /tmp read behavior. It is legitimate coordination guidance, but the unscoped pattern creates file-confusion risk.",
"severity": "low"
},
{
"id": "blocker:references/patterns.md:31:network-reconnaissance",
"verdict": "false_positive",
"confidence": 0.98,
"reason": "The phrase is part of a handoff checklist about current work status. It contains no network access or reconnaissance instruction."
},
{
"id": "external_commands:SKILL.md:40:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.98,
"reason": "The detected token is a markdown code fence around a natural-language prompt example. It is not Ruby or shell execution."
},
{
"id": "external_commands:SKILL.md:42:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.98,
"reason": "The detected token is a markdown code fence delimiter. The surrounding content is a prompt example, not executable code."
},
{
"id": "external_commands:SKILL.md:44:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.98,
"reason": "This is another markdown code fence used to display a prompt. It does not invoke a command interpreter."
},
{
"id": "external_commands:SKILL.md:46:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.98,
"reason": "The backticks close a markdown prompt example. No command, interpolation, or external execution is present."
},
{
"id": "external_commands:SKILL.md:50:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.98,
"reason": "The detected code fence wraps a natural-language context checkpoint prompt. It is documentation formatting only."
},
{
"id": "external_commands:SKILL.md:52:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.97,
"reason": "The region contains markdown guidance and reference text, not shell commands or Ruby backtick interpolation. No external command execution is instructed."
}
],
"semantic_findings": [
{
"title": "Credential Storage Guidance",
"description": "memory-management.md lists API endpoints and credentials as information to preserve. Persistent context should not store secrets because later sessions or agents may expose them.",
"severity": "medium",
"locations": [
{
"file": "memory-management.md",
"line_start": 54,
"line_end": 59
}
],
"confidence": 0.86,
"confidence_reasoning": "The text explicitly recommends preserving credentials in memory-related context. The risk is secret retention rather than code execution, so medium severity is appropriate."
}
],
"critical_findings": [],
"high_findings": [],
"medium_findings": [
{
"title": "Temp directory access",
"description": "output_path = f\"/tmp/agent_{task_id}_{timestamp}.md\"",
"locations": [
{
"file": "references/patterns.md",
"line_start": 50,
"line_end": 50
}
],
"review_kind": "capability",
"source_category": "filesystem",
"source_severity": "medium",
"confidence": 0.68,
"confidence_reasoning": "The example tells agents to write artifacts under predictable /tmp paths. That can collide with other files or expose coordination data on shared systems."
},
{
"title": "Temp directory access",
"description": "artifacts = glob.glob(\"/tmp/agent_*.md\")",
"locations": [
{
"file": "references/patterns.md",
"line_start": 53,
"line_end": 53
}
],
"review_kind": "capability",
"source_category": "filesystem",
"source_severity": "medium",
"confidence": 0.72,
"confidence_reasoning": "The example reads every matching /tmp artifact. Broad temp-directory reads can ingest stale or attacker-created files."
},
{
"title": "Credential Storage Guidance",
"description": "memory-management.md lists API endpoints and credentials as information to preserve. Persistent context should not store secrets because later sessions or agents may expose them.",
"locations": [
{
"file": "memory-management.md",
"line_start": 54,
"line_end": 59
}
],
"review_kind": "security",
"source_category": "semantic",
"source_severity": "medium",
"confidence": 0.86,
"confidence_reasoning": "The text explicitly recommends preserving credentials in memory-related context. The risk is secret retention rather than code execution, so medium severity is appropriate."
}
],
"low_findings": [
{
"title": "Python glob/pattern matching",
"description": "artifacts = glob.glob(\"/tmp/agent_*.md\")",
"locations": [
{
"file": "references/patterns.md",
"line_start": 53,
"line_end": 53
}
],
"review_kind": "capability",
"source_category": "filesystem",
"source_severity": "low",
"confidence": 0.7,
"confidence_reasoning": "The glob pattern supports the same broad /tmp read behavior. It is legitimate coordination guidance, but the unscoped pattern creates file-confusion risk."
}
],
"dangerous_patterns": [],
"files_scanned": 3,
"total_lines": 540,
"audit_model": "codex",
"audited_at": "2026-07-05T11:36:30.822Z",
"agent_auto_install_policy": "allowed",
"manual_install_policy": "allowed"
},
"content": {
"user_title": "Manage Multi-Agent Context",
"value_statement": "Long-running AI work often loses decisions, blockers, and session state. This skill provides practical workflows for checkpoints, handoffs, memory records, and compact context sharing.",
"seo_keywords": [
"Claude",
"Codex",
"Claude Code",
"context management",
"multi-agent workflows",
"AI handoff templates",
"session continuity",
"memory management",
"project knowledge base",
"context compression"
],
"actual_capabilities": [
"Captures key decisions, rationale, blockers, and active work for later sessions.",
"Prepares short or full handoff context for another agent or workflow stage.",
"Provides patterns for file-based context artifacts and coordination notes.",
"Guides persistent memory organization with entities, observations, and relationships.",
"Supports context compression by removing redundant details and preserving key facts.",
"Offers templates for completion and blocked-task handoffs."
],
"limitations": [
"Does not execute coordination automatically or create files without agent action.",
"Requires the user or agent to decide what information is safe to store.",
"Memory workflows depend on a compatible memory tool or manual tracking process.",
"Examples are guidance patterns, not tested automation scripts."
],
"use_cases": [
{
"title": "Coordinate Agent Handoffs",
"description": "Prepare focused context for the next agent, including current state, decisions, blockers, and expected next actions.",
"target_user": "AI workflow lead"
},
{
"title": "Resume Long Projects",
"description": "Create checkpoints that help a developer continue work after breaks, session changes, or task interruptions.",
"target_user": "software developer"
},
{
"title": "Maintain Project Memory",
"description": "Organize project knowledge with entities, observations, and relationships so important decisions remain searchable.",
"target_user": "technical team"
}
],
"prompt_templates": [
{
"title": "Capture Current Context",
"prompt": "Capture the current project context. Include the active goal, recent decisions, blockers, files involved, and next actions.",
"scenario": "Beginner checkpoint for a single active task."
},
{
"title": "Prepare Agent Handoff",
"prompt": "Prepare a handoff for the next agent. Include only context relevant to its task, expected output, and known constraints.",
"scenario": "Intermediate handoff for a multi-agent workflow."
},
{
"title": "Compress Session History",
"prompt": "Compress this session into a short context summary. Preserve decisions, rationale, unresolved questions, and risks.",
"scenario": "Advanced summary when token budget is limited."
},
{
"title": "Build Memory Graph Plan",
"prompt": "Design a memory graph plan for this project. Define entities, observations, relationship types, update rules, and safe storage boundaries.",
"scenario": "Advanced persistent memory setup for long-running work."
}
],
"output_examples": [
{
"input": "Capture context before this coding session ends.",
"output": "A concise checkpoint with the current goal, completed work, open decisions, blockers, and next actions."
},
{
"input": "Prepare a handoff for a reviewer agent.",
"output": "A reviewer-focused brief with changed areas, design choices, test status, risks, and files to inspect."
},
{
"input": "Compress our project memory for a new session.",
"output": "A short project memory summary that keeps durable facts and removes resolved discussion details."
}
],
"best_practices": [
"Store decisions with the reason, date, and affected project area.",
"Redact credentials and personal data before saving context.",
"Use short handoffs for routine transitions and full context for complex work."
],
"anti_patterns": [
"Saving full conversation transcripts when a short checkpoint is enough.",
"Recording credentials, tokens, or private data in persistent memory.",
"Passing every detail to each agent instead of tailoring context to the task."
],
"faq": [
{
"question": "What does this skill help manage?",
"answer": "It helps manage context across agents, sessions, checkpoints, and persistent project memory."
},
{
"question": "Does it run automation by itself?",
"answer": "No. It provides workflows, templates, and organization patterns for an agent or user to apply."
},
{
"question": "Can it help with Claude and Codex workflows?",
"answer": "Yes. The guidance is written for AI work in Claude, Codex, and Claude Code."
},
{
"question": "What should not be stored in memory?",
"answer": "Do not store credentials, tokens, private personal data, or secrets in persistent context."
},
{
"question": "When should I use a short context summary?",
"answer": "Use it for routine handoffs when the next agent only needs goals, decisions, blockers, and next steps."
},
{
"question": "When should I use full context?",
"answer": "Use full context for complex work that needs architecture, dependencies, decisions, and active work streams."
}
]
},
"file_structure": [
{
"name": "memory-management.md",
"type": "file",
"path": "memory-management.md",
"lines": 312
},
{
"name": "references",
"type": "dir",
"path": "references",
"children": [
{
"name": "patterns.md",
"type": "file",
"path": "references/patterns.md",
"lines": 131
}
]
},
{
"name": "SKILL.md",
"type": "file",
"path": "SKILL.md",
"lines": 97
}
]
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Unverified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |