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.
git clone --depth 1 https://github.com/FerroxLabs/ijfwWrote 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/commands/ferroxlabs/ijfw/memory-why)<a href="https://agentmods.dev/commands/ferroxlabs/ijfw/memory-why"><img src="https://agentmods.dev/badge/commands/ferroxlabs/ijfw/memory-why/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/commands/ferroxlabs/ijfw/memory-why"><img src="https://agentmods.dev/badge/commands/ferroxlabs/ijfw/memory-why.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 | $0.00018 | $0.00191 |
| Opus 5.5 | $0.00007 | $0.00076 |
| Sonnet 5.5 | $0.00004 | $0.00038 |
| Haiku 4.5 | $0.00002 | $0.00019 |
Grade A, and why
memory-why 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
For a given recall, report which signals contributed. Agent prompt:
When the last ijfw_memory_recall / ijfw_memory_search call returned a hit, surface:
<match-id>
BM25: <score> (matched terms: <tokens>)
vector: <cosine> (model: <model or "disabled">)
combined: <merged score>
snippet: <first 160 chars>
If vectors are disabled (IJFW_VECTORS=off or library not installed): say "Vectors disabled; result from keyword match only." rather than showing a 0 cosine that looks negative.
Use this to debug why a recall did (or didn't) find something the user expected.
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 First seen · 21 lines · 18 tokens per session scan A 5940638ef618
memory-why is a command published in the GitHub repository FerroxLabs/ijfw (214 stars, last pushed 5d ago), licensed MIT. It adds 18 tokens to every session and 191 once invoked, about $0.0001 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 2026-08-30.
Other commands, from other repositories
compact-task-memory
Produce a lossy compaction of TASKSTATE.md when task memory outgrows a useful size, keeping canonical decisions and the next step, dropping stale facts. Use when memory spans 5+ completed slices and feels heavy, resume cost grows, known facts are mostly resolved, or before a session pause. Do not use in early…
capture-observation
Append one observation, question, hypothesis, or concern to TASKSTATE.md as lean, append-only task memory, leaving other artifacts and work undisturbed. Use when something surfaces mid-work, a hypothesis needs logging, or a concern is too small for decision-interview or targeted-questions. Do not use for a decision…
init-workspace-flow-questions
Phase 8 Questions of init-workspace-flow.
code-locate
Locate code paths and line ranges for a described behavior: up to 10 candidates with HIGH/MEDIUM/LOW confidence and a search trail, plus a proposed SOURCEOFTRUTH.md update. Use when the first question is where one specific behavior lives and paths are unknown. Do not use when paths are explicit, the codebase is tiny…
_registry-protocol
This protocol is MANDATORY for ALL commands, agents, and phases.
harvest-session-learnings
Scan the session and task artifacts for reusable lessons (what was tried, what failed and why, what surprised us) and propose entries to append to LEARNINGS.md, the produce-side counterpart to what task-init reads. Use on demand mid-task after a hard-won fix or surprising failure, or at closure to sweep a long…