Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add rules/fmarzochi/egc/egc-contextgit clone --depth 1 https://github.com/Fmarzochi/EGCWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00364 | $0.00364 |
| Opus 5 | $0.00182 | $0.00182 |
| Sonnet 5 | $0.00073 | $0.00073 |
| Haiku 4.5 | $0.00036 | $0.00036 |
Grade A, and why
egc-context scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
EGC Project Memory
EGC Natural Language Interface
Detect user intent in any language and call the matching EGC tool — no keywords required:
Session
- User resumes work (any language) →
get_state - User ends session (any language) →
update_state
Diagnosis — when AI seems confused or hallucinating
- User questions whether things are working →
get_project_state - User asks what mistakes keep repeating →
detect_patterns - User asks what was learned in past sessions →
lesson_recall
Memory — user forces a save
- User asks to record a decision →
store_decision - User asks AI not to repeat a mistake →
lesson_save - User confirms a past lesson happened again →
lesson_reinforce - User wants to store something temporarily →
working_memory_set - User asks what is in temporary memory →
working_memory_get/working_memory_list
Search — when AI forgot something
- User asks about past decisions on a topic →
search_history - User asks for recent decisions chronologically →
query_history
Context — when heavy
- User says context is full or heavy →
reduce_context - User asks to compress session observations →
compress_observations
Safety — when user is suspicious
- User asks if a shell command is safe →
validate_command - User asks if a file path is safe to write →
validate_write - User asks to organize a complex task →
orchestrate_task - User asks AI to learn from session errors →
auto_learn
EGC Project Memory
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 3d ago First seen · 46 lines · 364 tokens per session scan A 76626605e1f0
egc-context is a cursor rule published in the GitHub repository Fmarzochi/EGC (46 stars, last pushed 7d ago), licensed Apache-2.0. It adds 364 tokens to every session, about $0.0018 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other cursor rules, from other repositories
cursorrules
Cursor rule "cursorrules" from mrzroot/agentforge, covering cursor ai project rules - generated by agentforge, tech stack, core directives, architectural standards and security & input validation.
project-rules
Global project rules and architecture standards for agentforge.
demand-discovery
Niche demand discovery research workflow — finds indie developer opportunities from 7 data sources.
archcore-files
Enforce MCP-only operations when working with .archcore/ files.
frontend-architecture
Vite + React SPA architecture - directory layout, providers, bundle splitting. Tailwind styling in tailwind.mdc.
database-design
Database design guidelines — relational normalisation, indexing, constraints, NoSQL patterns, query optimisation, migrations, security, and backup strategies. Applied when working with database schemas and queries.