1# Skills Mapping: X-to-Book System
2 
3This document provides a detailed mapping between the Agent Skills for Context Engineering and the design decisions in the X-to-Book system PRD.
4 
5## Skill: multi-agent-patterns
6 
7### Concepts Applied
8 
9| Concept | Skill Reference | PRD Application |
10|---------|-----------------|-----------------|
11| Supervisor pattern | "The supervisor pattern places a central agent in control, delegating to specialists and synthesizing results." | Orchestrator agent coordinates Scraper, Analyzer, Synthesizer, Writer, Editor agents |
12| Context isolation | "Sub-agents exist primarily to isolate context, not to anthropomorphize role division." | Each agent operates in clean context focused on its phase |
13| Telephone game problem | "LangGraph benchmarks found supervisor architectures initially performed 50% worse due to the 'telephone game' problem where supervisors paraphrase sub-agent responses incorrectly." | Phase outputs stored in file system, not passed through Orchestrator for synthesis |
14| File system coordination | "For complex tasks requiring shared state, agents read and write to persistent storage." | All inter-agent data flows through file system |
15| Supervisor bottleneck mitigation | "Implement output schema constraints so workers return only distilled summaries." | Orchestrator receives phase summaries, never raw data |
16 
17### Pattern Selection Rationale
18 
19The skill describes three patterns:
20 
211. **Supervisor/Orchestrator**: "When to use: Complex tasks with clear decomposition, tasks requiring coordination across domains."
222. **Peer-to-Peer/Swarm**: "When to use: Tasks requiring flexible exploration, tasks where rigid planning is counterproductive."
233. **Hierarchical**: "When to use: Large-scale projects with clear hierarchical structure."
24 
25**Selected**: Supervisor/Orchestrator
26 
27**Rationale**: Book production has clear sequential phases (scrape → analyze → synthesize → write → edit). Quality gates between phases require central coordination. Human oversight is important for content quality.
28 
29---
30 
31## Skill: context-fundamentals
32 
33### Concepts Applied
34 
35| Concept | Skill Reference | PRD Application |
36|---------|-----------------|-----------------|
37| Context as finite resource | "Context must be treated as a finite resource with diminishing marginal returns." | Explicit token budgets per agent (Orchestrator 50k, Writer 80k, etc.) |
38| Progressive disclosure | "Progressive disclosure manages context efficiently by loading information only as needed." | Book outline loads first; chapter content loads only when Writer works on that chapter |
39| Attention budget | "Models develop attention patterns from training data distributions where shorter sequences predominate." | Context limits set conservatively below model maximums |
40| Tool output volume | "Tool outputs comprise the majority of tokens in typical agent trajectories, with research showing observations can reach 83.9% of total context usage." | Tweet data processed separately, never enters main agent contexts |
41 
42### Context Budget Allocation
43 
44From skill: "Design with explicit context budgets in mind. Know the effective context limit for your model and task."
45 
46PRD implementation:
47 
48```yaml
49context_limits:
50  orchestrator: 50000   # Routing only, no raw data
51  scraper: 20000        # One account at a time
52  analyzer: 80000       # Pattern extraction
53  synthesizer: 100000   # Cross-account synthesis
54  writer: 80000         # Per-chapter drafting
55  editor: 60000         # Per-chapter review
56```
57 
58---
59 
60## Skill: memory-systems
61 
62### Concepts Applied
63 
64| Concept | Skill Reference | PRD Application |
65|---------|-----------------|-----------------|
66| Vector store limitations | "Vector stores lose relationship information... cannot answer 'What products did customers who purchased Product Y also buy?'" | Selected knowledge graph for relationship queries between accounts |
67| Temporal validity | "Temporal knowledge graphs add validity periods to facts. Each fact has a 'valid from' and optionally 'valid until' timestamp." | All relationships have temporal validity for tracking position evolution |
68| Entity memory | "Entity memory specifically tracks information about entities to maintain consistency." | Account, Tweet, Theme, Book, Chapter entity types defined |
69 
70### Memory Architecture Decision
71 
72From skill: "Choose memory architecture based on requirements: Simple persistence needs → File-system memory; Semantic search needs → Vector RAG; Relationship reasoning needs → Knowledge graph; Temporal validity needs → Temporal knowledge graph."
73 
74**Query requirements identified**:
75- "What has @account said about AI in the last 30 days?" → Temporal + entity filtering
76- "Which accounts disagree on crypto?" → Relationship traversal
77- "How has @account's position evolved?" → Temporal queries
78 
79**Selected**: Temporal Knowledge Graph
80 
81---
82 
83## Skill: context-optimization
84 
85### Concepts Applied
86 
87| Concept | Skill Reference | PRD Application |
88|---------|-----------------|-----------------|
89| Observation masking | "Observation masking replaces verbose tool outputs with compact references." | Raw tweet data stored in file system, not passed through context |
90| Compaction triggers | "Trigger compaction after significant memory accumulation, when retrieval returns too many outdated results." | 70% context utilization triggers compaction |
91| KV-cache optimization | "Place stable elements first (system prompt, tool definitions), then frequently reused elements, then unique elements last." | Context ordering: system prompt → tools → account config → daily outline → current task |
92 
93### Optimization Strategy
94 
95From skill: "When to optimize: Context utilization exceeds 70%, Response quality degrades as conversations extend."
96 
97PRD implementation:
98```python
99COMPACTION_THRESHOLD = 0.7  # 70% context utilization
100 
101if context_utilization > COMPACTION_THRESHOLD:
102    phase_outputs = compact_phase_outputs(phase_outputs)
103```
104 
105From skill: "What to apply: Tool outputs dominate → observation masking"
106 
107PRD implementation: All raw tweet data (potentially 100k+ tokens/day) is masked by:
1081. Scraper writes to file system
1092. Analyzer reads from file system, produces summaries
1103. Summaries (not raw data) flow to subsequent phases
111 
112---
113 
114## Skill: tool-design
115 
116### Concepts Applied
117 
118| Concept | Skill Reference | PRD Application |
119|---------|-----------------|-----------------|
120| Consolidation principle | "If a human engineer cannot definitively say which tool should be used in a given situation, an agent cannot be expected to do better." | 3 consolidated tools instead of 15+ narrow tools |
121| Description structure | "Effective tool descriptions answer four questions: What does the tool do? When should it be used? What inputs does it accept? What does it return?" | All tools have explicit usage triggers and error recovery |
122| Response format options | "Implementing response format options gives agents control over verbosity." | Tools support "concise" and "detailed" format parameters |
123| Error message design | "Error messages must be actionable. They must tell the agent what went wrong and how to correct it." | Errors include recovery guidance (RATE_LIMITED includes retry_after) |
124 
125### Tool Consolidation
126 
127From skill: "Instead of implementing list_users, list_events, and create_event, implement schedule_event that handles the full workflow internally."
128 
129PRD implementation:
130 
131**Before consolidation** (what we avoided):
132- `fetch_timeline`
133- `fetch_thread`
134- `fetch_engagement`
135- `search_tweets`
136- `store_entity`
137- `query_entities`
138- `update_validity`
139- etc.
140 
141**After consolidation**:
142- `x_data_tool` - all X data operations
143- `memory_tool` - all knowledge graph operations
144- `writing_tool` - all content operations
145 
146---
147 
148## Skill: evaluation
149 
150### Concepts Applied
151 
152| Concept | Skill Reference | PRD Application |
153|---------|-----------------|-----------------|
154| Multi-dimensional rubrics | "Agent quality is not a single dimension. It includes factual accuracy, completeness, coherence, tool efficiency, and process quality." | 5 weighted dimensions: Source Accuracy, Thematic Coherence, Completeness, Insight Quality, Readability |
155| LLM-as-judge | "LLM-based evaluation scales to large test sets and provides consistent judgments." | Automated evaluation for coherence and insight quality |
156| Human evaluation | "Human evaluation catches what automation misses." | Trigger human review when score < 0.7 or source accuracy < 0.8 |
157| Outcome-focused evaluation | "The solution is outcome-focused evaluation that judges whether agents achieve right outcomes while following reasonable processes." | Evaluate final book quality, not intermediate steps |
158 
159### Evaluation Rubric
160 
161From skill: "Effective rubrics cover key dimensions with descriptive levels."
162 
163PRD implementation:
164 
165| Dimension | Weight | Measurement |
166|-----------|--------|-------------|
167| Source Accuracy | 30% | Automated quote verification against original tweets |
168| Thematic Coherence | 25% | LLM-as-judge for narrative flow |
169| Completeness | 20% | Theme coverage calculation |
170| Insight Quality | 15% | LLM-as-judge for synthesis beyond restating |
171| Readability | 10% | Automated metrics + LLM judge |
172 
173---
174 
175## Cross-Skill Integration
176 
177The skills are designed to work together. This example demonstrates integration patterns:
178 
179| Integration | Skills Combined | Application |
180|-------------|-----------------|-------------|
181| Agent context budgets | multi-agent-patterns + context-fundamentals | Each agent has explicit limits based on role |
182| File system coordination | multi-agent-patterns + context-optimization | Avoids context passing, enables masking |
183| Memory-aware synthesis | memory-systems + context-optimization | Query relevant facts without loading full history |
184| Quality-driven routing | evaluation + multi-agent-patterns | Orchestrator uses quality scores for phase gates |
185 
186This integration is the core value proposition of the skills collection: they provide complementary patterns that address different aspects of context engineering while working together cohesively.
187 
188