Context Footprint
Approximate token footprint for each skill and agent when loaded into context.
How It Works
When you invoke a skill (e.g., /cf-plan), its full SKILL.md file is injected into the conversation prompt. This is the skill's context footprint — the amount of your context window it occupies just by being loaded.
The context footprint measures only the tokens inserted into the prompt when a skill or agent is loaded. It does not reflect the total tokens consumed during execution, which depends on how much code is read, how many tool calls are made, and how long the conversation runs.
Bootstrap context (~2,200 tokens) is loaded every session automatically. This includes the list of available skills, agents, activation signals, and conventions.
Agents run in forked sessions with their own context window, so they don't consume your main conversation's context.
Tier System
Each skill and agent is classified into a tier based on the approximate token count of its definition:
| Tier | Icon | Token Range | Meaning |
|---|---|---|---|
| Low | < 1,500 tokens | Lightweight — small prompt footprint | |
| Medium | 1,500 – 3,000 tokens | Moderate — standard prompt footprint | |
| High | > 3,000 tokens | Heavy — large prompt footprint |
Skills
Slash Commands
| Command | Status | Context | Approx. Tokens | Description |
|---|---|---|---|---|
| /cf-plan | updated | High> 3,000 tokens injected into prompt | ~8,000 | Brainstorm and plan |
| /cf-fix | updated | High> 3,000 tokens injected into prompt | ~3,800 | Quick bug fix workflow |
| /cf-review | updated | High> 3,000 tokens injected into prompt | ~3,700 | Multi-layer code review |
| /cf-scan | updated | High> 3,000 tokens injected into prompt | ~3,500 | Scan project and bootstrap memory |
| /cf-research | updated | High> 3,000 tokens injected into prompt | ~3,300 | In-depth research |
| /cf-ask | updated | High> 3,000 tokens injected into prompt | ~3,100 | Quick Q&A about codebase |
| /cf-help | updated | High> 3,000 tokens injected into prompt | ~3,100 | Answer questions about Coding Friend |
| /cf-remember | updated | High> 3,000 tokens injected into prompt | ~3,000 | Capture project knowledge |
| /cf-warm | betaupdated | Medium1,500 – 3,000 tokens injected into prompt | ~2,800 | Catch up after absence — git history summary |
| /cf-learn | updated | Medium1,500 – 3,000 tokens injected into prompt | ~2,700 | Extract human learning docs |
| /cf-review-in | updated | Medium1,500 – 3,000 tokens injected into prompt | ~2,600 | Collect external review results |
| /cf-teach | updated | Medium1,500 – 3,000 tokens injected into prompt | ~2,500 | Personal teacher — conversational storytelling breakdown |
| /cf-design | beta | Medium1,500 – 3,000 tokens injected into prompt | ~2,500 | UI design: scan patterns, design or modify UI consistently |
| /cf-optimize | updated | Medium1,500 – 3,000 tokens injected into prompt | ~2,500 | Structured optimization |
| /cf-review-out | updated | Low< 1,500 tokens injected into prompt | ~1,300 | Generate review prompt for external AI |
| /cf-session | betaupdated | Low< 1,500 tokens injected into prompt | ~1,200 | Save/load sessions |
| /cf-commit | updated | Low< 1,500 tokens injected into prompt | ~1,100 | Smart conventional commits |
| /cf-ship | updated | Low< 1,500 tokens injected into prompt | ~908 | Verify, commit, push, PR |
Auto-Invoked Skills
| Skill | Status | Context | Approx. Tokens | Activates When |
|---|---|---|---|---|
| cf-sys-debug | updated | High> 3,000 tokens injected into prompt | ~3,500 | Debugging issues |
| cf-tdd | updated | Medium1,500 – 3,000 tokens injected into prompt | ~2,100 | Writing new code (direct mode or TDD) |
| cf-verification | updated | Low< 1,500 tokens injected into prompt | ~637 | Before claiming done |
Agents
Agents run in forked sessions — they get their own context window and don't consume your main conversation's context. Token counts here refer to the size of the agent's system prompt injected into its forked context.
| Agent | Status | Context | Approx. Tokens | Model | Purpose |
|---|---|---|---|---|---|
cf-planner | Medium1,500 – 3,000 tokens injected into prompt | ~1,700 | Opus | Task decomposition | |
cf-reviewer | Medium1,500 – 3,000 tokens injected into prompt | ~1,500 | inherit | Review orchestrator | |
cf-explorer | Low< 1,500 tokens injected into prompt | ~1,400 | Haiku | Codebase exploration | |
cf-implementer | Low< 1,500 tokens injected into prompt | ~1,400 | Sonnet | Direct or TDD implementation | |
cf-writer | Low< 1,500 tokens injected into prompt | ~1,100 | Haiku | Lightweight doc writing | |
cf-reviewer-quality | Low< 1,500 tokens injected into prompt | ~1,100 | Haiku | Code quality + slop detection | |
cf-writer-deep | Low< 1,500 tokens injected into prompt | ~1,000 | Sonnet | Deep reasoning docs | |
cf-reviewer-security | Low< 1,500 tokens injected into prompt | ~961 | Sonnet | Security review (OWASP) | |
cf-reviewer-reducer | Low< 1,500 tokens injected into prompt | ~775 | Haiku | Deduplicate and rank findings | |
cf-reviewer-tests | Low< 1,500 tokens injected into prompt | ~707 | Haiku | Test coverage check | |
cf-reviewer-rules | Low< 1,500 tokens injected into prompt | ~682 | Haiku | Project rules compliance | |
cf-reviewer-plan | updated | Low< 1,500 tokens injected into prompt | ~670 | Sonnet | Plan alignment check |
Notes
- Token counts are approximate, calculated using
@lenml/tokenizer-claude. Actual Anthropic token counts may vary by ~20-30%. - Tier classification is what matters — small variations in exact count don't change the tier.
- These numbers reflect only the prompt injection size (the skill/agent definition loaded into context), not the total tokens used during execution.
- Custom skill guides (
.coding-friend/skills/<name>-custom/SKILL.md) add to the skill's prompt footprint when loaded. - Bootstrap context (~2,200 tokens) is always present regardless of which skill you use.