memory

A curated record of lessons from earlier work sessions in a code repository. Its entries guide future work but do not override fixed rules.

In plain words
What is it for?
Use it to record recurring problems, such as copied documentation drifting from source files or status claims becoming outdated.
Why use it?
It helps prevent repeated mistakes and keeps advice tied to what actually happened in earlier reviews.

Agent

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.

agentmods
npx agentmods add agents/adeelahmad/agentrc/memory
Clone the repo
git clone --depth 1 https://github.com/adeelahmad/agentrc
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 842 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00000 $0.00842
Opus 5 $0.00000 $0.00421
Sonnet 5 $0.00000 $0.00168
Haiku 4.5 $0.00000 $0.00084

Measured yesterday against content hash 7a9b7480bb5b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

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 yesterday.

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.

docs/agents/memory.md · 23 lines

How it starts

The opening of the file, as written. The whole thing — 23 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Memory — agentrc

Cross-sprint learning, curated at each retrospective. Entries are advisory and never relax an invariant. First agentic-agile session in this repo (Sprint 1); seeded from the 2026-07-02/03 docs+examples+spec audit that produced this work order.

Entries

id applies_to text origin recurrence added
M-001 all Inline doc/HTML snippets drift from the source-of-truth file they mirror (e.g. the "hello" Agentfile rendered on homepage/quickstart/examples-index diverged from examples/Agentfile.minimal). When a snippet duplicates a canonical file, its acceptance check MUST diff against that file, not just grep for a substring. 2026-07-02 audit; T1/T2/T4 3 2026-07-03
M-002 all Status/claim prose goes stale silently ("not yet published", "coming soon", present-tense "nothing to install") after the underlying fact changes. Re-derive every status claim from current ground truth (CLI code, published artifacts) rather than carrying prior wording forward. 2026-07-02 audit; T3 2 2026-07-03
M-003 all Locate every edit target by grep FIRST — repo layout is not assumed; the same string (e.g. tools/ping, IDENTITY name=hello) recurs across .md, .html, and Agentfile surfaces, and a partial sweep leaves the site internally inconsistent. work-order §0.9; T1/T12 2 2026-07-03
M-004 all On a static-site repo (Jekyll), the pipeline's own docs/agents/ planning artifacts carry YAML front matter and WILL be built into the site (22 junk pages + sitemap URLs leaked to prod). Add the pipeline artifact dir to the generator's exclude: at sprint start, and grep _site/live sitemap for it in the release checks. sprint1 T14 defect; PR #17 2 2026-07-03
M-005 all Never run supervisor git / working-tree commands (checkout, pull, append+commit) while a worker is active in the SHARED working tree — their branch/index ops collide (a supervisor git checkout failed mid-worker). Serialized-branch execution means: dispatch, WAIT for return, THEN do git. sprint1 wave-3 1 2026-07-03
M-006 all Live checks that grep rendered HTML for a multi-token literal (e.g. FROM python:3.11-slim) false-negative because syntax highlighting splits it across <span>s. Verify source for exact strings; for live, strip tags or check tokens (>FROM< + python:3.11-slim). sprint1 T14 1 2026-07-03
M-007 all In a set -o pipefail script, echo "$big" | grep -q PAT FALSE-negatives when the input exceeds the ~64KB pipe buffer AND the match is early: grep -q exits on first match, SIGPIPEs the echo, and pipefail reports the pipeline as failed even though it matched (a late-only match won't trip it — echo finishes first). Use here-strings grep -q PAT <<<"$big" (no pipe → no SIGPIPE); grep -c is immune (reads to EOF). Cost me a long T27 false-fail on the 97KB /spec/ page. sprint2 T27 1 2026-07-03

Read the full file on GitHub · 23 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. yesterday First seen · 23 lines · 0 tokens per session scan A 7a9b7480bb5b

Subscribe to this mod's changes

memory is an agent published in the GitHub repository adeelahmad/agentrc (4 stars, last pushed 1mo ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 842 tokens. 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 agents, from other repositories