Skills DirectorySkills Directory
SkillsLearnSecurityCategoriesDocsCommunityBlog
Sign InSubmit Skill
Skills Directory

Security-tested agent skills for Claude, coding agents, and AI workflows.

Directory

  • Browse Skills
  • All Skills A–Z
  • Claude Skills
  • Claude Code Skills
  • Agent Skills
  • Categories
  • Submit a Skill

Learn

  • Learn Hub
  • Install Claude Skills
  • Write SKILL.md
  • Skills vs MCP
  • Directories Compared

Security

  • Security
  • Methodology
  • Secure Claude Skills
  • Security Badges

Company

  • About
  • Community
  • Blog
  • API Docs
  • Advertise

2026 Skills Directory. All rights reserved.

Back to skills

Fidelis Memory

ASecurity

Use when a turn refers to earlier work, a past decision, a named project, or prior context that is not in the transcript — recall it from the local fidelis store before answering, and write the decision back after the turn settles. Also use when asked to check what is already known about something, or to record a durable fact. Requires a running fidelis-server; the plugin's bundled MCP server provides fidelis_orient, fidelis_recall, fidelis_query, and fidelis_health.

22 stars
0 votes
0 copies
1 views
Added 9/20/2026
ai-agentsgobashgitapi

Works with

cliapimcp

Security Analysis

A100/100

Scanned 9/20/2026

Install to Claude Code

$npx -y skills add hermes-labs-ai/fidelis --skill fidelis-memory --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Fidelis Memory?

Add the live security badge to your README — it updates automatically with every re-scan.

Security grade badge for Fidelis Memory
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/hermes-labs-ai-fidelis-memory/badge)](https://www.skillsdirectory.com/skills/hermes-labs-ai-fidelis-memory)

More formats (shields.io, HTML) on the badges page.

Download Zip
Files
SKILL.md
---
name: fidelis-memory
description: Use when a turn refers to earlier work, a past decision, a named project, or prior context that is not in the transcript — recall it from the local fidelis store before answering, and write the decision back after the turn settles. Also use when asked to check what is already known about something, or to record a durable fact. Requires a running fidelis-server; the plugin's bundled MCP server provides fidelis_orient, fidelis_recall, fidelis_query, and fidelis_health.
license: MIT
---

# fidelis memory workflow

This skill is the procedure for using the `fidelis` MCP tools that ship with
this plugin. It does not describe how fidelis is built; see
<https://github.com/hermes-labs-ai/fidelis> for that.

The retrieval path is zero-LLM by default: the server returns stored passages
verbatim and never regenerates them. Treat what comes back as a quotation, not
as a paraphrase you may polish.

## Before you start

Call `fidelis_health` once per session. It is the only cheap way to tell the
three failure modes apart:

| Result | Meaning | What to do |
|---|---|---|
| `status: ok \| memories: N` with N > 0 | Store is live and populated | Proceed |
| `status: ok \| memories: 0` | Server is up, store is empty | Say so; do not present an empty recall as "nothing was decided" |
| `fidelis-server unreachable at …` | No server on the configured port | Stop using these tools and tell the user to run `fidelis init` (or `fidelis server`) |

An unreachable server is not an empty memory. Never report the two the same way.

## Recall before answering

When the turn leans on something outside the transcript, retrieve first.

1. `fidelis_orient` — the default for a conversational turn. Pass the user's
   utterance verbatim as `utterance`, plus up to four `recent_turns` when the
   referent is a pronoun ("that project", "the thing we changed"). It returns
   an evidence-bound orientation packet or explicitly abstains. **An abstention
   is an answer**: it means the store has nothing bound to that referent. Say
   that instead of guessing.
2. `fidelis_recall` — when you already know what you are looking for and want
   ranked passages. Two-stage; zero-LLM unless a filter endpoint is configured.
3. `fidelis_query` — a narrow vector-only lookup. Use it to confirm whether one
   specific string is present, not to survey a topic.

Quote recalled text as stored. If a recalled memory contradicts the current
turn, surface the contradiction rather than silently preferring either side —
the store records what was decided, not what is still true.

## Store after a decision

Write back when the turn produced something a later session would have to
re-derive: a decision and its reason, a constraint, a rejected option and why
it was rejected, a coordinate (a path, an ID, a version) that was expensive to
find.

Writes go over the local HTTP API, which the MCP surface does not expose — it
is read-only by design:

```bash
curl -s -X POST http://127.0.0.1:19420/store \
  -H 'Content-Type: application/json' \
  -d '{"text": "Chose uvx over a repo checkout for the plugin MCP command so install needs no clone. 2026-09-18."}'
```

Write one self-contained fact per call, in past tense, with the date. A memory
that only makes sense next to the message that produced it is not worth
storing — the retrieval path returns the passage alone.

Do not store secrets, credentials, or anything the user marked as sensitive.
The store is plaintext on local disk.

## Don't

- Don't call `fidelis_recall` on every turn. Recall when the turn reaches
  outside the transcript; skip it for self-contained work.
- Don't rewrite recalled text into your own words and present it as stored.
- Don't treat a score as confidence. It ranks candidates against each other
  within one query; it is not calibrated across queries.
- Don't assume `/store` reached the vector store. When the embedder is
  unreachable the server answers `{"status": "queued"}` and replays later —
  read the response.

## Configuration

`FIDELIS_PORT` (client side) points the MCP server at a non-default
fidelis-server port. Set it in this plugin's `.mcp.json` `env` block and start
the server on the matching port.

Attribution

hermes-labs-aihermes-labs-ai
View sourceMore from hermes-labs-ai →
SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments (0)

No comments yet. Be the first to comment!

SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Related Skills

Caveman

Ultra-compressed communication mode that cuts output tokens while keeping technical accuracy. Levels: lite, full, ultra and the wenyan variants. Use for /caveman, "caveman mode", "talk like caveman", "be brief" or "less tokens".

1066601 votes

Hyperplan

Adversarial multi-agent planning skill. Self-orchestrates 5 hostile category members (unspecified-low, unspecified-high, deep, ultrabrain, artistry) via team-mode for ruthless cross-critique debate, distills only the defensible insights, then MANDATORILY hands the distilled insight bundle to the `plan` agent for executable plan formalization. Use when planning needs maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggers: 'hyperplan', 'hpp', '/hyperplan', ...

686011 votes

Mcp Code Execution

Routes multi-tool workflows through MCP servers for large datasets and pipelines. Use when Bash tool overhead is limiting throughput on data-heavy tasks.

3351 votes

catchup

Recovers the conversation and failed tool calls of a previous Codex, Claude Code, Antigravity, Cline, Copilot CLI, Cursor, DeepSeek Harness, Kimi, OpenCode, Pi Agent, or ZCode session. Use when the user says "catch up", "what did the last session do", "get me up to speed", "I switched agents", asks to recover/summarize a previous session before continuing, or asks to diagnose or report a catchup failure. Do NOT use for the current conversation, git history, or any non-agent log.

651 votes

math-skill

A comprehensive mathematical reasoning skill for AI assistants — handles arithmetic to research-level problems with rigorous step-by-step reasoning, systematic verification, and transparent uncertainty handling

381 votes
View all in ai-agents →