persistent-persona

persistent-persona is a skill for Claude Code, Codex from frankaging/lite-research-agents. It costs 35 tokens per session (860 once invoked), scanned A, original, MIT.

A mechanism for keeping an agent’s caller-defined persona consistent across sessions. It uses private memory to detect drift and encourages updating beliefs only when new evidence supports it.

In plain words
What is it for?
Maintaining a stable role or set of behavioral rules over time. Recording session summaries and checking whether changes in position are supported by new evidence.
Why use it?
Agents can change their stated position because of repetition, pressure, or group agreement rather than evidence. This helps distinguish genuine learning from unwanted persona drift.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: mentions subagents.

Good fit Maintaining a stable role or set of behavioral rules over time. Recording session summaries and checking whether changes in position are supported by new evidence.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/frankaging/lite-research-agents/persistent-persona
Install

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.

Any agent
npx skills add frankaging/lite-research-agents --skill persistent-persona
Clone the repo
git clone --depth 1 https://github.com/frankaging/lite-research-agents

Made for: Claude Code, Codex.

Wrote 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.

agentmods badge for persistent-persona

README.md
[![agentmods](https://agentmods.dev/badge/skills/frankaging/lite-research-agents/persistent-persona.svg)](https://agentmods.dev/skills/frankaging/lite-research-agents/persistent-persona)
Your own site
<a href="https://agentmods.dev/skills/frankaging/lite-research-agents/persistent-persona"><img src="https://agentmods.dev/badge/skills/frankaging/lite-research-agents/persistent-persona.svg" alt="Measured on agentmods" height="20"></a>
Per session 35 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 860 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00035 $0.00860
Opus 5 $0.00017 $0.00430
Sonnet 5 $0.00007 $0.00172
Haiku 4.5 $0.00003 $0.00086

Measured 7d ago against content hash 9e6a83218440, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade A, and why

persistent-persona 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 7d 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.

Origin

Copies of this mod

1 near-identical copy found in the catalogue:

skills/persistent-persona/SKILL.md · 105 lines

How it starts

The opening of the file, as written. The whole thing — 105 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Persistent Persona

This skill defines how to maintain a stable persona across sessions. It does not define what personas exist — those are defined by the caller (e.g. the war-room skill or the project brief).


Core Rule

A persona can be convinced by evidence. It cannot be convinced by pressure, repetition, or authority.

If another agent makes a good argument backed by new evidence, update your position. If the argument just gets louder, or the group converges — hold. Ask yourself: what actually changed? If the answer is nothing, do not update.


Biplate Memory

Each agent maintains a private memory file never shared with other agents:

memory/.private/agent-<index>-memo.md

This file is not referenced in SUMMARY.md, not included in any shared checkpoint folder, and not readable by any other agent.

What to Write

Append a new entry after every session. Do not overwrite prior entries — the full history reveals drift patterns over time.

## <YYYYMMDD_HHMMSS>

**Persona**: <persona name as defined by the caller>
**Session summary**: <one sentence — what did I contribute?>
**Position changes**: <did I update any positions? what evidence caused it?>
**Pressure events**: <did I feel pressure to drift? from whom? did I hold?>
**Mental state**: <am I still reasoning from my core stance, or mirroring the group?>
**Drift flag**: <yes | no>

Set Drift flag: yes if any of the following are true:

  • Agreed with the group without being shown new evidence
  • Softened a position to avoid conflict
  • Stopped asking the questions my persona would normally ask
  • My outputs sound more like another agent's persona than my own

Session Bootstrap

In agentic mode (war-room): the orchestrator runs steps 1–3 on the persona's behalf before building the subagent SI. The persona subagent receives the result already injected — it does not read its own memo directly.

Run this sequence before engaging:

1. Read memory/.private/agent-<index>-memo.md
2. Check the most recent Drift flag
3. If drift detected → inject persona reset (see below)
4. Follow memory-checkpoint load sequence:
     a. Read memory/SUMMARY.md
     b. Read latest memory/checkpoints/<timestamp>/agent-<index>.md
5. Re-read your persona definition as provided by the caller
6. Engage

Read the full file on GitHub · 105 lines

Changes

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.

  1. 7d ago First seen · 105 lines · 35 tokens per session scan A 9e6a83218440

Subscribe to this mod's changes

persistent-persona is a skill published in the GitHub repository frankaging/lite-research-agents (3 stars, last pushed 5mo ago), licensed MIT. It adds 35 tokens to every session and 860 once invoked, about $0.0002 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 2026-08-31.

Related

Other skills, from other repositories

autocontext-consumer

Use when an agent needs to USE knowledge Autocontext already produced - find which scenarios have knowledge, read the playbook and lessons for one, understand the on-disk file and folder layout, and move knowledge between checkouts. Host-agnostic; requires only the autoctx CLI and the filesystem.

greyhaven-ai/autocontext · 67 tokens

agent-memory-mcp

A hybrid memory system that provides persistent, searchable knowledge management for AI agents (Architecture, Patterns, Decisions).

lingxling/awesome-skills-cn · 26 tokens

log-session

Append a structured entry to the project's session log (docs/LOGS.md): what was done this session, files touched, decisions taken, and the next step — so the next session (or another person) can pick up the thread without re-reading git history. Run it before /clear, before closing Claude Code, or at any natural…

gtrabanco/agentic-workflow · 183 tokens

dpf-elicit-tacit-knowledge

Use when durable DPF knowledge lives only in a human's head — a decision rationale, the why behind a choice, a profession technique, domain context a build depends on.

OpenDigitalProductFactory/opendigitalproductfactory · 44 tokens

do-it-context

Use when project terms, relationships, or invariants drift between user language, docs, and implementation.

tdwhere123/do-it · 24 tokens

dpf-route-learning-to-commons

Use in the DPF codebase at a task or session boundary when a finding has been confirmed and is durable.

OpenDigitalProductFactory/opendigitalproductfactory · 31 tokens