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 skills add hermes-labs-ai/fidelis --skill fidelis-memorygit clone --depth 1 https://github.com/hermes-labs-ai/fidelisWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/hermes-labs-ai/fidelis/fidelis-memory)<a href="https://agentmods.dev/skills/hermes-labs-ai/fidelis/fidelis-memory"><img src="https://agentmods.dev/badge/skills/hermes-labs-ai/fidelis/fidelis-memory/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/hermes-labs-ai/fidelis/fidelis-memory"><img src="https://agentmods.dev/badge/skills/hermes-labs-ai/fidelis/fidelis-memory.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What 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 5.1 | $0.00063 | $0.00408 |
| Opus 5.5 | $0.00025 | $0.00163 |
| Sonnet 5.5 | $0.00013 | $0.00082 |
| Haiku 4.5 | $0.00006 | $0.00041 |
Grade A, and why
fidelis-memory 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 1mo 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
Fidelis memory workflow
Use fidelis_recall when the current turn needs earlier evidence. Default fast
recall uses local embeddings without a generative LLM. Use as_of for historical
validity and mode: "thorough" for explicit hybrid retrieval.
Read the record ID and temporal status alongside the text. Superseded records remain visible as history. Quote the original wording and preserve constraints; retrieval does not authenticate a claim or prove completeness. Similarity is a ranking signal, not calibrated confidence. Empty results do not prove absence.
Use fidelis_get to fetch full text and correction links by ID. Use
fidelis_recent to browse recent records or corrections. On errors, inspect
fidelis_health; an unavailable or unloaded store is not an empty one. Ask the
user to start the local service and Ollama when needed.
When the user intends a durable fact to be kept, call fidelis_store. To replace
an earlier statement, call fidelis_correct with its ID and the replacement
text. The original remains in history. Optional validity dates describe when
the statement applies. Read the acknowledgement: queued is not stored, duplicate
creates no new fact, and rejected writes were not accepted. Do not invent success.
Do not store secrets or content the user has excluded. Storage is local and not application-encrypted. Do not call recall for self-contained turns.
FIDELIS_PORT must match on the service and MCP process. Restart the client after
upgrading to refresh its six-tool surface.
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.
- 1mo ago Changed 748421d0b004
- 1mo ago Changed · -54 lines · -43 tokens per session cd8785a0b906
- 2mo ago First seen · 88 lines · 106 tokens per session scan A 14275dfb901e
fidelis-memory is a skill published in the GitHub repository hermes-labs-ai/fidelis (25 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 63 tokens to every session and 408 once invoked, about $0.0003 per session on Opus 5.5. 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 2026-09-19.
fidelis-memory compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| fidelis-memoryhermes-labs-ai/fidelis | Skill | 25 | 1mo ago | Apache-2.0 | A | 63 | Claude Code |
| agent-v3-memory-specialistruvnet/ruflo | Skill | 82k | yesterday | MIT | A | 25 | Claude Code |
| xerjxerj-org/xerj | Skill | 19k | today | Apache-2.0 | A | 38 | Claude Code, Codex |
| knowledge-orgbojieli/ai-agent-book | Skill | 59k | 1mo ago | Apache-2.0 | A | 96 | Claude Code, Codex |
| memory-to-skillzilliztech/memsearch | Skill | 2.9k | 5d ago | MIT | A | 82 | Claude Code, Codex |
| memory-recallzilliztech/memsearch | Skill | 2.9k | 5d ago | MIT | A | 152 | Claude Code |
| install-openviking-memoryvolcengine/OpenViking | Skill | 42k | yesterday | AGPL-3.0 | A | 191 | Claude Code, Codex |